- 15 Feb, 2018 4 commits
-
-
David S. Miller authored
Kirill Tkhai says: ==================== net: Add ioctl() SIOCGSKNS cmd to allow obtaining net ns of tun device Currently, it's not possible to get or check net namespace, which was used to create tun socket. User may have two tun devices with the same names in different nets, and there is no way to differ them each other. The patchset adds support for ioctl() cmd SIOCGSKNS for tun devices. It will allow people to obtain net namespace file descriptor like we allow to do that for sockets in general. v2: Add new patch [2/3] to export open_related_ns(). ==================== Signed-off-by:
Kirill Tkhai <ktkhai@virtuozzo.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Kirill Tkhai authored
This patch adds possibility to get tun device's net namespace fd in the same way we allow to do that for sockets. Socket ioctl numbers do not intersect with tun-specific, and there is already SIOCSIFHWADDR used in tun code. So, SIOCGSKNS number is choosen instead of custom-made for this functionality. Note, that open_related_ns() uses plain get_net_ns() and it's safe (net can't be already dead at this moment): tun socket is allocated via sk_alloc() with zero last arg (kern = 0). So, each alive socket increments net::count, and the socket is definitely alive during ioctl syscall. Also, common variable net is introduced, so small cleanup in TUNSETIFF is made. Signed-off-by:
-
Kirill Tkhai authored
This function will be used to obtain net of tun device. Signed-off-by:
-
Kirill Tkhai authored
This function will be used to obtain net of tun device. Signed-off-by:
-
- 14 Feb, 2018 36 commits
-
-
git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next-queueDavid S. Miller authored
Jeff Kirsher says: ==================== 40GbE Intel Wired LAN Driver Updates 2018-02-14 This patch series enables the new mqprio hardware offload mechanism creating traffic classes on VFs for XL710 devices. The parameters needed to configure these traffic classes/queue channels are provides by the user via the tc tool. A maximum of four traffic classes can be created on each VF. This patch series also enables application of cloud filters to each of these traffic classes. The cloud filters are applied using the tc-flower classifier. Example: 1. tc qdisc add dev vf0 root mqprio num_tc 4 map 0 0 0 0 1 2 2 3\ queues 2@0 2@2 1@4 1@5 hw 1 mode channel 2. tc qdisc add dev vf0 ingress 3. ethtool -K vf0 hw-tc-offload on 4. ip link set eth0 vf 0 spoofchk off 5. tc filter add dev vf0 protocol ip parent ffff: prio 1 flower dst_ip\ 192.168.3.5/32 ip_proto udp dst_port 25 skip_sw hw_tc 2 ==================== Signed-off-by: -
Tom Herbert authored
In kcm_attach strp_done is called when sk_user_data is already set to fail the attach. strp_done needs the strp to be stopped and warns if it isn't. Call strp_stop in this case to eliminate the warning message. Reported-by: syzbot+88dfb55e4c8b770d86e3@syzkaller.appspotmail.com Fixes: e5571240 ("kcm: Check if sk_user_data already set in kcm_attach" Signed-off-by:
Tom Herbert <tom@quantonium.net> Signed-off-by:
-
Wadim Egorov authored
Add documentation of ti,clk-output-sel which can be used to select a specific clock for CLK_OUT. Signed-off-by:
Wadim Egorov <w.egorov@phytec.de> Signed-off-by:
Daniel Schultz <d.schultz@phytec.de> Reviewed-by:
Florian Fainelli <f.fainelli@gmail.com> Signed-off-by:
-
Wadim Egorov authored
The DP83867 has a muxing option for the CLK_OUT pin. It is possible to set CLK_OUT for different channels. Create a binding to select a specific clock for CLK_OUT pin. Signed-off-by:
Andrew Lunn <andrew@lunn.ch> Reviewed-by:
-
Jon Maloy authored
Currently, the default link tolerance set in struct tipc_bearer only has effect on links going up after that moment. I.e., a user has to reset all the node's links across that bearer to have the new value applied. This is too limiting and disturbing on a running cluster to be useful. We now change this so that also already existing links are updated dynamically, without any need for a reset, when the bearer value is changed. We leverage the already existing per-link functionality for this to achieve the wanted effect. Acked-by:
Ying Xue <ying.xue@windriver.com> Signed-off-by:
Jon Maloy <jon.maloy@ericsson.com> Signed-off-by:
-
David S. Miller authored
Rahul Lakkireddy says: ==================== cxgb4: speed up reading on-chip memory This series of patches speed up reading on-chip memory (EDC and MC) by reading 64-bits at a time. Patch 1 reworks logic to read EDC and MC. Patch 2 adds logic to read EDC and MC 64-bits at a time. v2: - Dropped AVX CPU intrinsic instructions. - Use readq() to read 64-bits at a time. ==================== Signed-off-by: -
Rahul Lakkireddy authored
Use readq() (via t4_read_reg64()) to read 64-bits at a time. Read residual in 32-bit multiples. Signed-off-by:
Rahul Lakkireddy <rahul.lakkireddy@chelsio.com> Signed-off-by:
Ganesh Goudar <ganeshgr@chelsio.com> Signed-off-by:
-
Rahul Lakkireddy authored
Rework logic to read EDC and MC. Do 32-bit reads at a time. Signed-off-by:
-
David Ahern authored
IPv4 uses set_lwt_redirect to set the lwtunnel redirect functions as needed. Move it to lwtunnel.h as lwtunnel_set_redirect and change IPv6 to also use it. Signed-off-by:
David Ahern <dsahern@gmail.com> Signed-off-by:
-
David S. Miller authored
Andrew Lunn says: ==================== PTP support for DSA and mv88e6xxx driver. This patchset adds support for using the PTP hardware in switches supported by the mv88e6xxx driver. The code was produces in collaboration with Brandon Streiff doing the initial implementation, and then Richard Cochran and Andrew Lunn making further changes and cleanups. The code is sufficient to use ptp4l on a single DSA interface, either as a master or a slave. Due to the use of an MDIO bus to access the switch, reading hardware timestamps is slower than what ptp4l expects. Thus it is necessary to use the option --tx_timestamp_timeout=32. Heavy use of ethtool -S, or bridge fdb show can also upset ptp4l. Patches to address this will follow. Further work is requires to support bridges using Boundary Clock or Transparent Clock mode. Since the RFC, an overflow bug has been fixed. Brandon Streiff has also Acked-by: the updates to his initial patchset. ==================== Signed-off-by: -
Brandon Streiff authored
88E6341 devices default to timestamping at the PHY, but due to a hardware issue, timestamps via this component are unreliable. For this family, configure the PTP hardware to force the timestamping to occur at the MAC. Signed-off-by:
Brandon Streiff <brandon.streiff@ni.com> Signed-off-by:
-
Brandon Streiff authored
This patch implements RX/TX timestamping support. The Marvell PTP hardware supports RX timestamping individual message types, but for simplicity we only support the EVENT receive filter since few if any clients bother with the more specific filter types. checkpatch and reverse Christmas tree changes by Andrew Lunn. Re-factor duplicated code paths and avoid IfOk anti-pattern, use the common ptp worker thread from the class layer and time stamp UDP/IPv4 frames as well as Layer-2 frame by Richard Cochran. Signed-off-by:
Richard Cochran <richardcochran@gmail.com> Signed-off-by:
-
Brandon Streiff authored
Forward the rx/tx timestamp machinery from the dsa infrastructure to the switch driver. On the rx side, defer delivery of skbs until we have an rx timestamp. This mimicks the behavior of skb_defer_rx_timestamp. On the tx side, identify PTP packets, clone them, and pass them to the underlying switch driver before we transmit. This mimicks the behavior of skb_tx_timestamp. Adjusted txstamp API to keep the allocation and freeing of the clone in the same central function by Richard Cochran Signed-off-by:
-
Brandon Streiff authored
This patch adds support to the dsa slave network device so that switch drivers can implement the SIOC[GS]HWTSTAMP ioctls and the ethtool timestamp-info interface. Signed-off-by:
-
Brandon Streiff authored
This patch adds support for configuring mv88e6xxx GPIO lines as PTP pins, so that they may be used for time stamping external events or for periodic output. Checkpatch and reverse Christmas tree fixes by Andrew Lunn Periodic output removed by Richard Cochran, until a better abstraction of a VCO is added to Linux in general. Signed-off-by:
-
Brandon Streiff authored
MV88E6352 and later switches support GPIO control through the "Scratch & Misc" global2 register. (Older switches do too, though with a slightly different register interface. Only the 6352-style is implemented here.) Add a new file, global2_scratch.c, for operations in the Scratch & Misc space. Additionally, add a GPIO operations structure to present an abstract view over GPIO manipulation. Reverse Christmas tree and unsigned has been replaced with unsigned int by Andrew Lunn. Signed-off-by:
-
Brandon Streiff authored
This patch adds basic support for exposing the 32-bit timestamp counter inside the mv88e6xxx switch as a ptp_clock. Adjfine implemented by Richard Cochran. Andrew Lunn: fix return value of PTP stub function. Signed-off-by:
-
Brandon Streiff authored
This patch implements support for accessing the Precision Time Protocol and Time Application Interface registers via the AVB register interface in the Global 2 register. The register interface differs slightly between different models; older models use a 3-bit operations field, while newer models use a 2-bit field. The operations values and the special "global port" values are different between the two. This is a similar split to the differences in the "Ingress Rate" register between models, so, like in that case, we call the two variants "6352" and "6390" and create an ops structure to abstract between the two. checkpatch fixups by Andrew Lunn Signed-off-by:
-
Brandon Streiff authored
Let the mv88e6xxx_g2_* register accessor functions be accessible outside of global2.c. Signed-off-by:
-
Andrew Lunn authored
When NET_PTP_CLASSIFY is disabled, a stub function is required in order that the drivers compile. Signed-off-by:
-
Eric Dumazet authored
배석진 reported that in some situations, packets for a given 5-tuple end up being processed by different CPUS. This involves RPS, and fragmentation. 배석진 is seeing packet drops when a SYN_RECV request socket is moved into ESTABLISH state. Other states are protected by socket lock. This is caused by a CPU losing the race, and simply not caring enough. Since this seems to occur frequently, we can do better and perform a second lookup. Note that all needed memory barriers are already in the existing code, thanks to the spin_lock()/spin_unlock() pair in inet_ehash_insert() and reqsk_put(). The second lookup must find the new socket, unless it has already been accepted and closed by another cpu. Note that the fragmentation could be avoided in the first place by use of a correct TCP MSS option in the SYN{ACK} packet, but this does not mean we can not be more robust. Many thanks to 배석진 for a very detailed analysis. Reported-by:배석진 <soukjin.bae@samsung.com> Signed-off-by:
Eric Dumazet <edumazet@google.com> Signed-off-by:
-
Avinash Dayanand authored
This patch provides support to add or delete cloud filter for queue channels created for ADq on VF. We are using the HW's cloud filter feature and programming it to act as a TC filter applied to a group of queues. There are two possible modes for a VF when applying a cloud filter 1. Basic Mode: Intended to apply filters that don't need a VF to be Trusted. This would include the following Dest MAC + L4 port Dest MAC + VLAN + L4 port 2. Advanced Mode: This mode is only for filters with combination that requires VF to be Trusted. Dest IP + L4 port When cloud filters are applied on a trusted VF and for some reason the same VF is later made as untrusted then all cloud filters will be deleted. All cloud filters has to be re-applied in such a case. Cloud filters are also deleted when queue channel is deleted. Testing-Hints: ============= 1. Adding Basic Mode filter should be possible on a VF in Non-Trusted mode. 2. In Advanced mode all filters should be able to be created. Steps: ====== 1. Enable ADq and create TCs using TC mqprio command 2. Apply cloud filter. 3. Turn-off the spoof check. 4. Pass traffic. Example: ======== 1. tc qdisc add dev enp4s2 root mqprio num_tc 4 map 0 0 0 0 1 2 2 3\ queues 2@0 2@2 1@4 1@5 hw 1 mode channel 2. tc qdisc add dev enp4s2 ingress 3. ethtool -K enp4s2 hw-tc-offload on 4. ip link set ens261f0 vf 0 spoofchk off 5. tc filter add dev enp4s2 protocol ip parent ffff: prio 1 flower\ dst_ip 192.168.3.5/32 ip_proto udp dst_port 25 skip_sw hw_tc 2 Signed-off-by:
Avinash Dayanand <avinash.dayanand@intel.com> Tested-by:
Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by:
Jeff Kirsher <jeffrey.t.kirsher@intel.com>
-
Harshitha Ramamurthy authored
This patch enables a tc filter to be applied as a cloud filter for the VF. This patch adds functions which parse the tc filter, extract the necessary fields needed to configure the filter and package them in a virtchnl message to be sent to the PF to apply them. Signed-off-by:
Harshitha Ramamurthy <harshitha.ramamurthy@intel.com> Tested-by:
-
Harshitha Ramamurthy authored
This patch adds infrastructure to send virtchnl messages to the PF to configure filters on the VF. The patch adds a struct called virtchnl_filter which contains information about the fields in the user-specified tc filter. Signed-off-by:
-
Harshitha Ramamurthy authored
This patch adds a macro to check if the size of a union is correct. It throws a divide by zero error if the union is not of the correct size. Signed-off-by:
-
Avinash Dayanand authored
This patch handles the request from ADq enabled VF to allocate bandwidth to each traffic class which means for each VSI. Signed-off-by:
-
Harshitha Ramamurthy authored
This patch adds support to configure bandwidth for the traffic classes via tc tool. The required information is passed to the PF which is used in the process of setting up the traffic classes. Signed-off-by:
-
Avinash Dayanand authored
This patch takes care of freeing up all the VSIs, queues and other ADq related software and hardware resources, when a user requests for deletion of ADq on VF. Example command: tc qdisc del dev eth0 root Signed-off-by:
-
Avinash Dayanand authored
This patch allocates number of queues requested by the user as a part of TC command when ADq is enabled on a VF. In order to be consistent in design with PF implementation of ADq, don't allow to set channels via ethtool from VF when ADq is already enabled. This means the users will not be able to change the number of queues/channels via ethtool for a VF when ADq is ON. In order to be able to use set channels, users will be required to disable ADq first and then try setting the channels again. When ADq is enabled on VF, it goes through a reset during which VSIs and queues are re-configured. Meanwhile if we receive link status message from PF even before the queues are re-configured, just ignore this link up message. Signed-off-by:
-
Avinash Dayanand authored
This patch enables ADq and creates queue channels on a VF. An ADq enabled VF can have up to 4 VSIs and each one of them represents a traffic class and this is termed as a queue channel. Each of these VSIs can have up to 4 queues. This patch services the request for enabling ADq and adds queue channel based on the TC mqprio info provided by the user in the VF. Initially a check is made to see if spoof check is OFF, if not ADq will not be enabled. PF notifies VF for a reset in order to complete the creation of ADq resources i.e. creation of additional VSIs and allocation of queues as per TC information, all in the reset path. Steps: ====== 1. Turn off the spoof check 2. Enable ADq using tc mqprio command with or without rate limit. 3. Pass traffic. Example: ======== % ip link set dev eth0 vf 0 spoofchk off % tc qdisc add dev $iface root mqprio num_tc 4 map\ 0 0 0 0 1 1 1 1 2 2 2 2 3 3 3 3 queues\ 4@0 4@4 4@8 4@8 hw 1 mode channel Expected results: ================= 1. Total number of queues for the VF should be sum of queues of all TCs. 2. Traffic flow should be normal without errors. Signed-off-by:
-
Harshitha Ramamurthy authored
This patch introduces the callback to the ndo_setup_tc function in the VF driver. We add a wrapper function to make room for the upcoming cloud filter patches which add calls to different functions from setup_tc. First, we add support for capability exchange for ADQ between the PF and VF. Next, we add support to take in the mqprio configuration and configure queues as per the traffic classes, rate limit and the priorities specified by the user. This is done by passing the channel config to the PF driver through a virtchannel message. The flags and bits added, track if ADq is enabled, set max number of traffic classes to 4 and provide ability to negotiate capability with the PF. Signed-off-by:
-
Harshitha Ramamurthy authored
This patch defines new structs in support of the virtchannel message that the VF sends to the PF to create a queue channel specified by the user via tc tool. Signed-off-by:
-
Avinash Dayanand authored
One of the previous patch fixes the link up issue by ignoring it if i40evf is not in __I40EVF_RUNNING state. However this doesn't fix the race condition when queues are disabled esp for ADq on VF. Hence check if all queues are enabled before starting all queues. Signed-off-by:
-
David S. Miller authored
David Ahern says: ==================== net: dev: Make protocol ptr dependent on CONFIG Found these in a branch from 3-years ago. Still relevant today. Make decnet, ax25, and atalk ptrs in net_device based on their respective CONFIG. ==================== Signed-off-by: -
David Ahern authored
Signed-off-by:
-
David Ahern authored
Signed-off-by:
David Ahern <dsa@cumulusnetworks.com> Signed-off-by:
-
