- 05 Dec, 2016 17 commits
-
-
Florian Fainelli authored
Implement ethtooll::nway_restart by utilizing mii_nway_restart. Signed-off-by:
Florian Fainelli <f.fainelli@gmail.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
David S. Miller authored
Merge branch 'for-upstream' of git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth-next Johan Hedberg says: ==================== pull request: bluetooth-next 2016-12-03 Here's a set of Bluetooth & 802.15.4 patches for net-next (i.e. 4.10 kernel): - Fix for a potential NULL deref in the ieee802154 netlink code - Fix for the ED values of the at86rf2xx driver - Documentation updates to ieee802154 - Cleanups to u8 vs __u8 usage - Timer API usage cleanups in HCI drivers Please let me know if there are any issues pulling. Thanks. ==================== Signed-off-by:
-
David S. Miller authored
Eric Dumazet says: ==================== tcp: tsq: performance series Under very high TX stress, CPU handling NIC TX completions can spend considerable amount of cycles handling TSQ (TCP Small Queues) logic. This patch series avoids some atomic operations, but most notable patch is the 3rd one, allowing other cpus processing ACK packets and calling tcp_write_xmit() to grab TCP_TSQ_DEFERRED so that tcp_tasklet_func() can skip already processed sockets. This avoid lots of lock acquisitions and cache lines accesses, particularly under load. In v2, I added : - tcp_small_queue_check() change to allow 1st and 2nd packets in write queue to be sent, even in the case TX completion of already acknowledged packets did not happen yet. This helps when TX completion coalescing parameters are set even to insane values, and/or busy polling is used. - A reorganization of struct sock fields to lower false sharing and increase data locality. - Then I moved tsq_flags from tcp_sock to struct sock also to reduce cache line misses during TX completions. I measured an overall throughput gain of 22 % for heavy TCP use over a single TX queue. ==================== Signed-off-by: -
Eric Dumazet authored
tsq_flags being in the same cache line than sk_wmem_alloc makes a lot of sense. Both fields are changed from tcp_wfree() and more generally by various TSQ related functions. Prior patch made room in struct sock and added sk_tsq_flags, this patch deletes tsq_flags from struct tcp_sock. Signed-off-by:
Eric Dumazet <edumazet@google.com> Signed-off-by:
-
Eric Dumazet authored
Group fields used in TX path, and keep some cache lines mostly read to permit sharing among cpus. Gained two 4 bytes holes on 64bit arches. Added a place holder for tcp tsq_flags, next to sk_wmem_alloc to speed up tcp_wfree() in the following patch. I have not added ____cacheline_aligned_in_smp, this might be done later. I prefer doing this once inet and tcp/udp sockets reorg is also done. Tested with both TCP and UDP. UDP receiver performance under flood increased by ~20 % : Accessing sk_filter/sk_wq/sk_napi_id no longer stalls because sk_drops was moved away from a critical cache line, now mostly read and shared. /* --- cacheline 4 boundary (256 bytes) --- */ unsigned int sk_napi_id; /* 0x100 0x4 */ int sk_rcvbuf; /* 0x104 0x4 */ struct sk_filter * sk_filter; /* 0x108 0x8 */ union { struct socket_wq * sk_wq; /* 0x8 */ struct socket_wq * sk_wq_raw; /* 0x8 */ }; /* 0x110 0x8 */ struct xfrm_policy * sk_policy[2]; /* 0x118 0x10 */ struct dst_entry * sk_rx_dst; /* 0x128 0x8 */ struct dst_entry * sk_dst_cache; /* 0x130 0x8 */ atomic_t sk_omem_alloc; /* 0x138 0x4 */ int sk_sndbuf; /* 0x13c 0x4 */ /* --- cacheline 5 boundary (320 bytes) --- */ int sk_wmem_queued; /* 0x140 0x4 */ atomic_t sk_wmem_alloc; /* 0x144 0x4 */ long unsigned int sk_tsq_flags; /* 0x148 0x8 */ struct sk_buff * sk_send_head; /* 0x150 0x8 */ struct sk_buff_head sk_write_queue; /* 0x158 0x18 */ __s32 sk_peek_off; /* 0x170 0x4 */ int sk_write_pending; /* 0x174 0x4 */ long int sk_sndtimeo; /* 0x178 0x8 */ Signed-off-by:Paolo Abeni <pabeni@redhat.com> Signed-off-by:
-
Eric Dumazet authored
Adding a likely() in tcp_mtu_probe() moves its code which used to be inlined in front of tcp_write_xmit() We still have a cache line miss to access icsk->icsk_mtup.enabled, we will probably have to reorganize fields to help data locality. Signed-off-by:
-
Eric Dumazet authored
Always allow the two first skbs in write queue to be sent, regardless of sk_wmem_alloc/sk_pacing_rate values. This helps a lot in situations where TX completions are delayed either because of driver latencies or softirq latencies. Test is done with no cache line misses. Signed-off-by:
-
Eric Dumazet authored
Under high load, tcp_wfree() has an atomic operation trying to schedule a tasklet over and over. We can schedule it only if our per cpu list was empty. Signed-off-by:
-
Eric Dumazet authored
Under high stress, I've seen tcp_tasklet_func() consuming ~700 usec, handling ~150 tcp sockets. By setting TCP_TSQ_DEFERRED in tcp_wfree(), we give a chance for other cpus/threads entering tcp_write_xmit() to grab it, allowing tcp_tasklet_func() to skip sockets that already did an xmit cycle. In the future, we might give to ACK processing an increased budget to reduce even more tcp_tasklet_func() amount of work. Signed-off-by:
-
Eric Dumazet authored
Instead of atomically clear TSQ_THROTTLED and atomically set TSQ_QUEUED bits, use one cmpxchg() to perform a single locked operation. Since the following patch will also set TCP_TSQ_DEFERRED here, this cmpxchg() will make this addition free. Signed-off-by:
-
Eric Dumazet authored
This is a cleanup, to ease code review of following patches. Old 'enum tsq_flags' is renamed, and a new enumeration is added with the flags used in cmpxchg() operations as opposed to single bit operations. Signed-off-by:
-
David S. Miller authored
Michael Chan says: ==================== bnxt_en: Add DCBNL support. This series adds DCBNL operations to support host-based IEEE DCBX. v2: Updated to the latest firmware interface spec. David, please consider this series for net-next. ==================== Signed-off-by: -
Michael Chan authored
Report PFC statistics to ethtool -S and DCBNL. Signed-off-by:
Michael Chan <michael.chan@broadcom.com> Signed-off-by:
-
Michael Chan authored
Support only IEEE DCBX initially. Add IEEE DCBNL ops and functions to get and set the hardware DCBX parameters. The DCB code is conditional on Kconfig CONFIG_BNXT_DCB. Signed-off-by:
-
Michael Chan authored
Latest interface has the latest DCB command structs. Get and store the max number of lossless TCs the hardware can support. Signed-off-by:
-
Michael Chan authored
Add a new function bnxt_setup_mq_tc() to handle MQPRIO. This new function will be called during ETS setup when we add DCBNL in the next patch. Signed-off-by:
-
Jesper Nilsson authored
According to the documentation, the PHYs supported by this driver can also support pause frames. Announce this to be so. Tested with a TI83822I. Acked-by:
Andrew F. Davis <afd@ti.com> Signed-off-by:
Jesper Nilsson <jesper.nilsson@axis.com> Signed-off-by:
-
- 04 Dec, 2016 23 commits
-
-
Erik Nordmark authored
Implemented RFC7527 Enhanced DAD. IPv6 duplicate address detection can fail if there is some temporary loopback of Ethernet frames. RFC7527 solves this by including a random nonce in the NS messages used for DAD, and if an NS is received with the same nonce it is assumed to be a looped back DAD probe and is ignored. RFC7527 is enabled by default. Can be disabled by setting both of conf/{all,interface}/enhanced_dad to zero. Signed-off-by:Erik Nordmark <nordmark@arista.com> Signed-off-by:
Bob Gilligan <gilligan@arista.com> Reviewed-by:
Hannes Frederic Sowa <hannes@stressinduktion.org> Signed-off-by:
-
David S. Miller authored
Andrew Lunn says: ==================== mv88e6390 batch 3 More patches to support the MV88e6390. This is mostly refactoring existing code and adding implementations for the mv88e6390. This patchset set which reserved frames are sent to the cpu, the size of jumbo frames that will be accepted, turn off egress rate limiting, and configuration of pause frames. ==================== Signed-off-by: -
Andrew Lunn authored
The mv88e6390 has a number flow control registers accessed via the Flow Control register. Use these to set the pause control. Signed-off-by:
Andrew Lunn <andrew@lunn.ch> Signed-off-by:
-
Andrew Lunn authored
The mv88e6390 has a different mechanism for configuring pause. Refactor the code into an ops function, and for the moment, don't add any mv88e6390 code yet. Signed-off-by:
-
Andrew Lunn authored
There are two different rate limiting configurations, depending on the switch generation. Refactor this into ops. Signed-off-by:
-
Andrew Lunn authored
Some switches support jumbo frames. Refactor this code into operations in the ops structure. Signed-off-by:
-
Andrew Lunn authored
Older devices have a couple of registers in global2. The mv88e6390 family has a single register in global1 behind which hides similar configuration. Implement and op for this. Signed-off-by:
-
David S. Miller authored
Andrew Lunn says: ==================== MV88E6390 batch two This is the second batch of patches adding support for the MV88e6390. They are not sufficient to make it work properly. The mv88e6390 has a much expanded set of priority maps. Refactor the existing code, and implement basic support for the new device. Similarly, the monitor control register has been reworked. The mv88e6390 has something odd in its EDSA tagging implementation, which means it is not possible to use it. So we need to use DSA tagging. This is the first device with EDSA support where we need to use DSA, and the code does not support this. So two patches refactor the existing code. The two different register definitions are separated out, and using DSA on an EDSA capable device is added. v2: Add port prefix Add helper function for 6390 Add _IEEE_ into #defines Split monitor_ctrl into a number of separate ops. Remove 6390 code which is management, used in a later patch s/EGREES/EGRESS/. Broke up setup_port_dsa() and set_port_dsa() into a number of ops v3: Verify mandatory ops for port setup Don't set ether type for DSA port. ==================== Signed-off-by: -
Andrew Lunn authored
Older chips only support DSA tagging. Newer chips have both DSA and EDSA tagging. Refactor the code by adding port functions for setting the frame mode, egress mode, and if to forward unknown frames. This results in the helper mv88e6xxx_6065_family() becoming unused, so remove it. Signed-off-by:
-
Andrew Lunn authored
Older chips support a single tagging protocol, DSA. New chips support both DSA and EDSA, an enhanced version. Having both as an option changes the register layouts. Up until now, it has been assumed that if EDSA is supported, it will be used. Hence the register layout has been determined by which protocol should be used. However, mv88e6390 has a different implementation of EDSA, which requires we need to use the DSA tagging. Hence separate the selection of the protocol from the register layout. Signed-off-by:
Vivien Didelot <vivien.didelot@savoirfairelinux.com> Signed-off-by:
-
Andrew Lunn authored
The mv88e6390 changes the monitor control register into the Monitor and Management control, which is an indirection register to various registers. Add ops to set the CPU port and the ingress/egress port for both register layouts, to global1 Signed-off-by:
-
Andrew Lunn authored
The mv88e6390 does not have the two registers to set the frame priority map. Instead it has an indirection registers for setting a number of different priority maps. Refactor the old code into an function, implement the mv88e6390 version, and use an op to call the right one. Signed-off-by:
-
David S. Miller authored
Jiri Pirko says: ==================== ipv4: fib: Replay events when registering FIB notifier Ido says: In kernel 4.9 the switchdev-specific FIB offload mechanism was replaced by a new FIB notification chain to which modules could register in order to be notified about the addition and deletion of FIB entries. The motivation for this change was that switchdev drivers need to be able to reflect the entire FIB table and not only FIBs configured on top of the port netdevs themselves. This is useful in case of in-band management. The fundamental problem with this approach is that upon registration listeners lose all the information previously sent in the chain and thus have an incomplete view of the FIB tables, which can result in packet loss. This patchset fixes that by dumping the FIB tables and replaying notifications previously sent in the chain for the registered notification block. The entire dump process is done under RCU and thus the FIB notification chain is converted to be atomic. The listeners are modified accordingly. This is done in the first eight patches. The ninth patch adds a change sequence counter to ensure the integrity of the FIB dump. The last patch adds the dump itself to the FIB chain registration function and modifies existing listeners to pass a callback to be executed in case dump was inconsistent. --- v3->v4: - Register the notification block after the dump and protect it using the change sequence counter (Hannes Frederic Sowa). - Since we now integrate the dump into the registration function, drop the sysctl to set maximum number of retries and instead set it to a fixed number. Lets see if it's really a problem before adding something we can never remove. - For the same reason, dump FIB tables for all net namespaces. - Add a comment regarding guarantees provided by mutex semantics. v2->v3: - Add sysctl to set the number of FIB dump retries (Hannes Frederic Sowa). - Read the sequence counter under RTNL to ensure synchronization between the dump process and other processes changing the routing tables (Hannes Frederic Sowa). - Pass a callback to the dump function to be executed prior to a retry. - Limit the dump to a single net namespace. v1->v2: - Add a sequence counter to ensure the integrity of the FIB dump (David S. Miller, Hannes Frederic Sowa). - Protect notifications from re-ordering in listeners by using an ordered workqueue (Hannes Frederic Sowa). - Introduce fib_info_hold() (Jiri Pirko). - Relieve rocker from the need to invoke the FIB dump by registering to the FIB notification chain prior to ports creation. ==================== Signed-off-by: -
Ido Schimmel authored
Commit b90eb754 ("fib: introduce FIB notification infrastructure") introduced a new notification chain to notify listeners (f.e., switchdev drivers) about addition and deletion of routes. However, upon registration to the chain the FIB tables can already be populated, which means potential listeners will have an incomplete view of the tables. Solve that by dumping the FIB tables and replaying the events to the passed notification block. The dump itself is done using RCU in order not to starve consumers that need RTNL to make progress. The integrity of the dump is ensured by reading the FIB change sequence counter before and after the dump under RTNL. This allows us to avoid the problematic situation in which the dumping process sends a ENTRY_ADD notification following ENTRY_DEL generated by another process holding RTNL. Callers of the registration function may pass a callback that is executed in case the dump was inconsistent with current FIB tables. The number of retries until a consistent dump is achieved is set to a fixed number to prevent callers from looping for long periods of time. In case current limit proves to be problematic in the future, it can be easily converted to be configurable using a sysctl. Signed-off-by:
Ido Schimmel <idosch@mellanox.com> Signed-off-by:
Jiri Pirko <jiri@mellanox.com> Signed-off-by:
-
Ido Schimmel authored
The next patch will enable listeners of the FIB notification chain to request a dump of the FIB tables. However, since RTNL isn't taken during the dump, it's possible for the FIB tables to change mid-dump, which will result in inconsistency between the listener's table and the kernel's. Allow listeners to know about changes that occurred mid-dump, by adding a change sequence counter to each net namespace. The counter is incremented just before a notification is sent in the FIB chain. Signed-off-by:
-
Ido Schimmel authored
In order not to hold RTNL for long periods of time we're going to dump the FIB tables using RCU. Convert the FIB notification chain to be atomic, as we can't block in RCU critical sections. Signed-off-by:
-
Ido Schimmel authored
We can miss FIB notifications sent between the time the ports were created and the FIB notification block registered. Instead of receiving these notifications only when they are replayed for the FIB notification block during registration, just register the notification block before the ports are created. Signed-off-by:
-
Ido Schimmel authored
Convert rocker to offload FIBs in deferred work in a similar fashion to mlxsw, which was converted in the previous commits. Signed-off-by:
-
Ido Schimmel authored
As explained in the previous commits, we need to process FIB entries addition / deletion events in FIFO order or otherwise we can have a mismatch between the kernel's FIB table and the device's. Create an ordered workqueue for rocker to which these work items will be submitted to. Signed-off-by:
-
Ido Schimmel authored
FIB offload is currently done in process context with RTNL held, but we're about to dump the FIB tables in RCU critical section, so we can no longer sleep. Instead, defer the operation to process context using deferred work. Make sure fib info isn't freed while the work is queued by taking a reference on it and releasing it after the operation is done. Deferring the operation is valid because the upper layers always assume the operation was successful. If it's not, then the driver-specific abort mechanism is called and all routed traffic is directed to slow path. The work items are submitted to an ordered workqueue to prevent a mismatch between the kernel's FIB table and the device's. Signed-off-by:
-
Ido Schimmel authored
We're going to start processing FIB entries addition / deletion events in deferred work. These work items must be processed in the order they were submitted or otherwise we can have differences between the kernel's FIB table and the device's. Solve this by creating an ordered workqueue to which these work items will be submitted to. Note that we can't simply convert the current workqueue to be ordered, as EMADs re-transmissions are also processed in deferred work. Later on, we can migrate other work items to this workqueue, such as FDB notification processing and nexthop resolution, since they all take the same lock anyway. Signed-off-by:
-
Ido Schimmel authored
As explained in the previous commit, modules are going to need to take a reference on fib info and then drop it using fib_info_put(). Add the fib_info_hold() helper to make the code more readable and also symmetric with fib_info_put(). Signed-off-by:
-
Ido Schimmel authored
The FIB notification chain is going to be converted to an atomic chain, which means switchdev drivers will have to offload FIB entries in deferred work, as hardware operations entail sleeping. However, while the work is queued fib info might be freed, so a reference must be taken. To release the reference (and potentially free the fib info) fib_info_put() will be called, which in turn calls free_fib_info(). Export free_fib_info() so that modules will be able to invoke fib_info_put(). Signed-off-by:
-
