- 25 Feb, 2019 32 commits
-
-
Jiri Pirko authored
Track the basic codepaths of delta rehash handling, using mlxsw tracepoints. Use IPv6 addresses. Signed-off-by:
Jiri Pirko <jiri@mellanox.com> Signed-off-by:
Ido Schimmel <idosch@mellanox.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Jiri Pirko authored
Other mutexes are taking care of proper locking for this, no longer needed to take RTNL mutex here. Signed-off-by:
-
Jiri Pirko authored
No longer require RTNL lock in this code. Newly introduced mutexes take care of guarding objagg and bloom filter. There is no need to guard gen_pool_alloc()/gen_pool_free() as they are fine to be called lockless. Signed-off-by:
-
Jiri Pirko authored
Relax dependency on rtnl mutex during vregion_rehash_intrvl_set(). The vregion list is protected with newly introduced mutex. Signed-off-by:
-
Jiri Pirko authored
Protect objagg structures by adding a mutex to ERP code and take it during the structure manipulation. Signed-off-by:
-
Jiri Pirko authored
For MR ACL profile is does not make sense to do periodical rehashes, as there is only one mask in use during the whole vregion lifetime. Therefore periodical work is scheduled but the rehash never happens. So allow to enable/disable rehash for the whole group, which is added per-profile. Disable rehashing for MR profile. Addition to the vregion list is done only in case the rehash is enable on the particular vregion. Also, the addition is moved after delayed work init to avoid schedule of uninitialized work from vregion_rehash_intrvl_set(). Symmetrically, deletion from the list is done before canceling the delayed work so it is not scheduled by vregion_rehash_intrvl_set() again. Signed-off-by:
-
Jiri Pirko authored
Bloom filter is shared within multiple regions. For updates, it needs to be guarded by a separate mutex. Do that in order to not rely on RTNL mutex. Signed-off-by:
-
Jiri Pirko authored
In order to remove dependency on RTNL, introduce a mutex to guard vregion structure, list of chunks and list of entries in chunks. Signed-off-by:
-
Jiri Pirko authored
Refactor existing _vchunk_assoc/_vchunk_deassoc() functions into _vregion_get()/_vregion_put() to make the code simpler and prepared for vregion locking. Signed-off-by:
-
Jiri Pirko authored
In order to remove RTNL lock dependency, it is needed to protect the regions list in a group. Introduce a mutex to do the job. Signed-off-by:
-
Jiri Pirko authored
Make the existing group structure to contain fields needed for HW region list manipulations. Move the rest of the fields into new vgroup struct. This makes layering cleaner as the vgroup struct is on higher level than low-level group struct. Also, this makes it possible to introduce fine-grained locking. Signed-off-by:
-
Jiri Pirko authored
Never used, remove it. Signed-off-by:
-
David S. Miller authored
Tristram Ha says: ==================== net: dsa: microchip: add MIB counters support This series of patches is to modify the KSZ9477 DSA driver to read MIB counters periodically to avoid overflow. The MIB counters should be read only when there is link. Otherwise it is a waste of time as hardware never increases the counters. Functions are added to check the port link status so that MIB counters read call is used efficiently. v4 - Use readx_poll_timeout - Fix using mutex in a timer callback function problem - use dp->slave directly instead of checking whether it is valid - Add port_cleanup function in a separate patch - Add a mutex so that changing device variables is safe v3 - Use netif_carrier_ok instead of checking the phy device pointer v2 - Create macro similar to readx_poll_timeout to use with switch - Create ksz_port_cleanup function so that variables like on_ports and live_ports can be updated inside it v1 - Use readx_poll_timeout - Do not clear MIB counters when port is enabled - Do not advertise 1000 half-duplex mode when port is enabled - Do not use freeze function as MIB counters may miss counts ==================== Signed-off-by: -
Tristram Ha authored
Add port_cleanup function to reset some device variables when the port is disabled. Add a mutex to make sure changing those variables is thread-safe. Signed-off-by:
Tristram Ha <Tristram.Ha@microchip.com> Reviewed-by:
Florian Fainelli <f.fainelli@gmail.com> Signed-off-by:
-
Tristram Ha authored
Remove unnecessary header include. Signed-off-by:
Andrew Lunn <andrew@lunn.ch> Reviewed-by:
-
Tristram Ha authored
Get port link status to know whether to read MIB counters when the link is going down. Signed-off-by:
-
Tristram Ha authored
Add background MIB counter reading support. Port MIB counters should only be read when there is link. Otherwise it is a waste of time as hardware never increases those counters. There are exceptions as some switches keep track of dropped counts no matter what. Signed-off-by:
-
Tristram Ha authored
Prepare PHY for proper advertisement as sometimes the PHY in the switch has its own problems even though it may share the PHY id from regular PHY but the fixes in the PHY driver do not apply. Signed-off-by:
-
David S. Miller authored
Maxime Chevallier says: ==================== net: phy: marvell10g: Add 2.5GBaseT support This series adds the missing bits necessary to fully support 2.5GBaseT in the Marvell Alaska PHYs. The main points for that support are : - Making use of the .get_features call, recently introduced by Heiner and Andrew, that allows having a fully populated list of supported modes, including 2500BaseT. - Configuring the MII to 2500BaseX when establishing a link at 2.5G - Adding a small quirk to take into account the fact that some PHYs in the family won't report the correct supported abilities The rest of the series consists of small cosmetic improvements such as using the correct helper to set a linkmode bit and adding macros for the PHY ids. We also add support for the 88E2110 PHY, which doesn't require the quirk, and support for 2500BaseT in the PPv2 driver, in order to have a fully working setup on the MacchiatoBin board. Changes since V1 : Fixed formatting issue in patch 01, rebased. ==================== Signed-off-by: -
Maxime Chevallier authored
This patch adds support for the 88x2110 PHY, which is similar to the already supported 88x3310 PHY without the SFP interface. It supports 10/100/1000BASET along with 2.5GBASET, 5GBASET and 10GBASET, with the same interface modes that are used by the 3310. This PHY don't have the same issue as the 88x3310 regarding 2.5/5G abilities, and correctly follows the 802.3bz standard to list the supported abilities. Signed-off-by:
Maxime Chevallier <maxime.chevallier@bootlin.com> Suggested-by:
Antoine Tenart <antoine.tenart@bootlin.com> Reviewed-by:
-
Maxime Chevallier authored
The PPv2 controller is able to support 2.5G speeds, allowing to use 2.5GBASET in conjunction with PHYs that use 2500BASEX as their MII interface when using this mode. Signed-off-by:
-
Maxime Chevallier authored
As per 802.3bz, if bit 14 of (1.11) "PMA Extended Abilities" indicates whether or not we should read register (1.21) "2.52/5G PMA Extended Abilities", which contains information on the support of 2.5GBASET and 5GBASET. After testing on several variants of PHYS of this family, it appears that bit 14 in (1.11) isn't always set when it should be. PHYs 88X3310 (on MacchiatoBin) and 88E2010 do support 2.5G and 5GBASET, but don't have 1.11.14 set. Their register 1.21 is filled with the correct values, indicating 2.5G and 5G support. PHYs 88E2110 do have their 1.11.14 bit set, as it should. Signed-off-by:
-
Maxime Chevallier authored
The PHY ID corresponding to the 88X3310 is also used for other PHYs in the same family, such as the 88E2010. Use a #define for the PHY id, that ignores the last nibble. Signed-off-by:
-
Maxime Chevallier authored
The Marvell Alaska family of PHYs supports 2.5GBaseT and 5GBaseT modes, as defined in the 802.3bz specification. Upon establishing a 2.5GBASET link, the PHY will reconfigure it's MII interface to 2500BASEX. At 5G, the PHY will reconfigure it's interface to 5GBASE-R, but this mode isn't supported by any MAC for now. This was tested with : - The 88X3310, which is on the MacchiatoBin - The 88E2010, an Alaska PHY that has no fiber interfaces, and is limited to 5G maximum speed. Signed-off-by:
-
Maxime Chevallier authored
Cosmetic patch making use of helpers dedicated to linkmodes handling. Signed-off-by:
-
Maxime Chevallier authored
The Alaska family of 10G PHYs has more abilities than the ones listed in PHY_10GBIT_FULL_FEATURES, the exact list depending on the model. Make use of the newly introduced .get_features call to build this list, using genphy_c45_pma_read_abilities to build the list of supported linkmodes, and adding autoneg ability based on what's reported by the AN MMD. .config_init is still used to validate the interface_mode. Signed-off-by:
-
Heiner Kallweit authored
The current code reports a link as up if all devices (except a few blacklisted ones) report the link as up. This breaks Aquantia AQCS109 for lower speeds because on this PHY the PCS link status reflects a 10G link only. For Marvell there's a similar issue, therefore PHYXS device isn't checked. There may be more PHYs where depending on the mode the link status of only selected devices is relevant. For now it seems to be sufficient to check the link status of the PMAPMD device only. Leave the loop in the code to be prepared in case we have to add functionality to check more than one device, depending on the mode. Successfully tested on a board with an AQCS109. Signed-off-by:
Heiner Kallweit <hkallweit1@gmail.com> Signed-off-by:
-
David S. Miller authored
Florian Fainelli says: ==================== net: switchdev.h inclusion removal This targets a few drivers that no longer to have net/switchdev.h included. ==================== Signed-off-by: -
Florian Fainelli authored
This is no longer necessary after eca59f69 ("net: Remove support for bridge bypass ndos from stacked devices") Suggested-by:
Andy Gospodarek <andy@greyhouse.net> Acked-by:
-
Florian Fainelli authored
This is no longer necessary after a5084bb7 ("nfp: Implement ndo_get_port_parent_id()") Signed-off-by:
Jakub Kicinski <jakub.kicinski@netronome.com> Acked-by:
-
Hauke Mehrtens authored
eth_change_mtu() is not needed any more, the networking subsystem will call it automatically when this callback is not implemented. Signed-off-by:
Hauke Mehrtens <hauke@hauke-m.de> Signed-off-by:
-
Heiner Kallweit authored
The way to define __ETHTOOL_LINK_MODE_MASK_NBITS seems to be overly complicated, go with a standard approach instead. Whilst we're at it, move the comment to the right place. v2: - rebased Signed-off-by:
-
- 24 Feb, 2019 8 commits
-
-
David S. Miller authored
Andy Roulin says: ==================== net: protodown support for macvlan and vxlan This patch series adds dev_change_proto_down_generic, a generic implementation of ndo_change_proto_down, which sets the netdev carrier state according to the new proto_down value. This handler adds the ability to set protodown on macvlan and vxlan interfaces in a generic way for use by control protocols like VRRPD. Patch (1) introduces the handler in net/code/dev.c. Patch (2) and (3) add support for change_proto_down in macvlan and vxlan drivers, respectively, using the new function. ==================== Signed-off-by: -
Andy Roulin authored
Add ndo_change_proto_down support through dev_change_proto_down_generic for use by control protocols like VRRPD. Signed-off-by:
Andy Roulin <aroulin@cumulusnetworks.com> Acked-by:
Roopa Prabhu <roopa@cumulusnetworks.com> Signed-off-by:
-
Andy Roulin authored
Add ndo_change_proto_down support through dev_change_proto_down_generic for use by control protocols like VRRPD. Signed-off-by:
-
Andy Roulin authored
Introduce dev_change_proto_down_generic, a generic ndo_change_proto_down implementation, which sets the netdev carrier state according to proto_down. This adds the ability to set protodown on vxlan and macvlan devices in a generic way for use by control protocols like VRRPD. Signed-off-by:
-
David S. Miller authored
Vlad Buslov says: ==================== Add tests for unlocked flower classifier implementation Implement tests for tdc testsuite to verify concurrent rules update with rtnl-unlocked flower classifier implementation. The goal of these tests is to verify general flower classifier correctness by updating filters on same classifier instance in parallel and to verify its atomicity by concurrently updating filters in same handle range. All three filter update operations (add, replace, delete) are tested. Existing script tdc_batch.py is re-used for batch file generation. It is extended with several optional CLI arguments that are needed for concurrency tests. Thin wrapper tdc_multibatch.py is implemented on top of tdc_batch.py to simplify its usage when generating multiple batch files for several test configurations. Parallelism in tests is implemented by running multiple instances of tc in batch mode with xargs tool. Xargs is chosen for its ease of use and because it is available by default on most modern Linux distributions. ==================== Acked-by:
Cong Wang <xiyou.wangcong@gmail.com> Signed-off-by:
-
Vlad Buslov authored
Implement test that runs 5 instances of tc replace filter in parallel with 5 instances of tc del filter from same tp instance. Each instance uses its own filter handle and key range. Signed-off-by:
Vlad Buslov <vladbu@mellanox.com> Signed-off-by:
-
Vlad Buslov authored
Implement test that runs 5 instances of tc add filter in parallel with 5 instances of tc del filter from same tp instance. Each instance uses its own filter handle and key range. Extend tdc_multibatch.py with additional options required to implement the test: common prefix for all generated batch files, first value of filter handle range, MAC address prefix modifier. These are necessary to allow creating batch files with unique keys and handle ranges with multiple invocation of tdc_multibatch.py helper script. Signed-off-by:
-
Vlad Buslov authored
Implement test that verifies concurrent deletion of rules by executing 10 tc instances that delete flower filters in same handle range. In this case only one tc instance succeeds in deleting a filter with particular handle. To mitigate expected failures of all other instances, run tc with 'force' option to continue processing batch file in case of errors and expect xargs to return code '123' that indicates that invocation of command(s) exited with error in range 1-125. Signed-off-by:
-
