- 09 Jul, 2020 17 commits
-
-
Danielle Ratson authored
Test port split configuration using previously added number of port lanes attribute. Check that all the splittable ports are successfully split to their maximum number of lanes and below, and that those which are not splittable fail to be split. Test output example: TEST: swp4 is unsplittable [ OK ] TEST: split port swp53 into 4 [ OK ] TEST: Unsplit port pci/0000:03:00.0/25 [ OK ] TEST: split port swp53 into 2 [ OK ] TEST: Unsplit port pci/0000:03:00.0/25 [ OK ] Signed-off-by:
Danielle Ratson <danieller@mellanox.com> Reviewed-by:
Petr Machata <petrm@mellanox.com> Signed-off-by:
Ido Schimmel <idosch@mellanox.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Danielle Ratson authored
Currently, all the input checks are done in driver. After adding the split capability to devlink port, move the checks to devlink. Signed-off-by:
Jiri Pirko <jiri@mellanox.com> Signed-off-by:
-
Danielle Ratson authored
Add a new attribute that indicates the split ability of devlink port. Drivers are expected to set it via devlink_port_attrs_set(), before registering the port. Signed-off-by:
-
Danielle Ratson authored
Currently, port attributes like flavour, port number and whether the port was split are set when initializing a port. Set the split ability of the port as well, based on port_mapping->width field and split attribute of devlink port in spectrum, so that it could be easily passed to devlink in the next patch. Signed-off-by:
-
Danielle Ratson authored
Add a new devlink port attribute that indicates the port's number of lanes. Drivers are expected to set it via devlink_port_attrs_set(), before registering the port. The attribute is not passed to user space in case the number of lanes is invalid (0). Signed-off-by:
-
Danielle Ratson authored
Currently, port attributes like flavour, port number and whether the port was split are set when initializing a port. Set the number of lanes of the port as well so that it could be easily passed to devlink in the next patch. Signed-off-by:
-
Danielle Ratson authored
Currently, devlink_port_attrs_set accepts a long list of parameters, that most of them are devlink port's attributes. Use the devlink_port_attrs struct to replace the relevant parameters. Signed-off-by:
-
Danielle Ratson authored
The struct devlink_port_attrs holds the attributes of devlink_port. Similarly to the previous patch, 'switch_port' attribute is another exception. Move 'switch_port' to be devlink_port's field. Signed-off-by:
-
Danielle Ratson authored
The struct devlink_port_attrs holds the attributes of devlink_port. The 'set' field is not devlink_port's attribute as opposed to most of the others. Move 'set' to be devlink_port's field called 'attrs_set'. Signed-off-by:
-
kernel test robot authored
drivers/net/phy/mscc/mscc_ptp.c:1496:1-3: WARNING: PTR_ERR_OR_ZERO can be used Use PTR_ERR_OR_ZERO rather than if(IS_ERR(...)) + PTR_ERR Generated by: scripts/coccinelle/api/ptr_ret.cocci Fixes: 7d272e63 ("net: phy: mscc: timestamping and PHC support") CC: Antoine Tenart <antoine.tenart@bootlin.com> Signed-off-by:
kernel test robot <lkp@intel.com> Signed-off-by:
-
Colin Ian King authored
Currently the u16 skb->vlan_tci is being right shifted twice by VLAN_PRIO_SHIFT, once in the macro skb_vlan_tag_get_pri and explicitly by VLAN_PRIO_SHIFT afterwards. The combined shift amount is larger than the u16 so the end result is always zero. Remove the second explicit shift as this is extraneous. Fixes: 6e9fdb60 ("net: systemport: Add support for VLAN transmit acceleration") Addresses-Coverity: ("Operands don't affect result") Signed-off-by:
Colin Ian King <colin.king@canonical.com> Acked-by:
Florian Fainelli <f.fainelli@gmail.com> Signed-off-by:
-
David S. Miller authored
Paolo Abeni says: ==================== mptcp: introduce msk diag interface This series implements the diag interface for the MPTCP sockets. Since the MPTCP protocol value can't be represented with the current diag uAPI, the first patch introduces an extended attribute allowing user-space to specify lager protocol values. The token APIs are then extended to allow traversing the whole token container. Patch 3 carries the actual diag interface implementation, and later patch bring-in some functional self-tests. ==================== Signed-off-by: -
Paolo Abeni authored
basic functional test, triggering the msk diag interface code. Require appropriate iproute2 support, skip elsewhere. Reviewed-by:
Mat Martineau <mathew.j.martineau@linux.intel.com> Signed-off-by:
Paolo Abeni <pabeni@redhat.com> Signed-off-by:
-
Paolo Abeni authored
exposes basic inet socket attribute, plus some MPTCP socket fields comprising PM status and MPTCP-level sequence numbers. Reviewed-by:
-
Paolo Abeni authored
mptcp_token_iter_next() allow traversing all the MPTCP sockets inside the token container belonging to the given network namespace with a quite standard iterator semantic. That will be used by the next patch, but keep the API generic, as we plan to use this later for PM's sake. Additionally export mptcp_token_get_sock(), as it also will be used by the diag module. Reviewed-by:
-
Paolo Abeni authored
After commit bf976514 ("sock: Make sk_protocol a 16-bit value") the current size of 'sdiag_protocol' is not sufficient to represent the possible protocol values. This change introduces a new inet diag request attribute to let user space specify the relevant protocol number using u32 values. The attribute is parsed by inet diag core on get/dump command and the extended protocol value, if available, is preferred to 'sdiag_protocol' to lookup the diag handler. The parse attributed are exposed to all the diag handlers via the cb->data. Note that inet_diag_dump_one_icsk() is left unmodified, as it will not be used by protocol using the extended attribute. Suggested-by:
Christoph Paasch <cpaasch@apple.com> Signed-off-by:
-
Xu Wang authored
This patch is to use eth_broadcast_addr() to assign broadcast address insetad of memset(). Signed-off-by:
Xu Wang <vulab@iscas.ac.cn> Signed-off-by:
-
- 08 Jul, 2020 23 commits
-
-
Jarod Wilson authored
The bond_ipsec_* helpers don't need RTNL, and can potentially get called without it being held, so switch from rtnl_dereference() to rcu_dereference() to access bond struct data. Lightly tested with xfrm bonding, no problems found, should address the syzkaller bug referenced below. Reported-by: syzbot+582c98032903dcc04816@syzkaller.appspotmail.com CC: Huy Nguyen <huyn@mellanox.com> CC: Saeed Mahameed <saeedm@mellanox.com> CC: Jay Vosburgh <j.vosburgh@gmail.com> CC: Veaceslav Falico <vfalico@gmail.com> CC: Andy Gospodarek <andy@greyhouse.net> CC: "David S. Miller" <davem@davemloft.net> CC: Jeff Kirsher <jeffrey.t.kirsher@intel.com> CC: Jakub Kicinski <kuba@kernel.org> CC: Steffen Klassert <steffen.klassert@secunet.com> CC: Herbert Xu <herbert@gondor.apana.org.au> CC: netdev@vger.kernel.org CC: intel-wired-lan@lists.osuosl.org Signed-off-by:
Jarod Wilson <jarod@redhat.com> Signed-off-by:
-
Fabio Estevam authored
DP83869 is an Ethernet PHY, not a charger, so fix the documentation accordingly. Fixes: 4d66c56f ("dt-bindings: net: dp83869: Add TI dp83869 phy") Signed-off-by:
Fabio Estevam <festevam@gmail.com> Acked-by:
Dan Murphy <dmurphy@ti.com> Signed-off-by:
-
Fabio Estevam authored
DP83867 is an Ethernet PHY, not a charger, so fix the documentation accordingly. Fixes: 74ac28f1 ("dt-bindings: dp83867: Convert DP83867 to yaml") Signed-off-by:
-
Jarod Wilson authored
It's possible that device removal happens when the bond is in non-AB mode, and addition happens in AB mode, so bond_ipsec_del_sa() never gets called, which leaves security associations in an odd state if bond_ipsec_add_sa() then gets called after switching the bond into AB. Just call add and delete universally for all modes to keep things consistent. However, it's also possible that this code gets called when the system is shutting down, and the xfrm subsystem has already been disconnected from the bond device, so we need to do some error-checking and bail, lest we hit a null ptr deref. Fixes: a3b658cf ("bonding: allow xfrm offload setup post-module-load") CC: Huy Nguyen <huyn@mellanox.com> CC: Saeed Mahameed <saeedm@mellanox.com> CC: Jay Vosburgh <j.vosburgh@gmail.com> CC: Veaceslav Falico <vfalico@gmail.com> CC: Andy Gospodarek <andy@greyhouse.net> CC: "David S. Miller" <davem@davemloft.net> CC: Jeff Kirsher <jeffrey.t.kirsher@intel.com> CC: Jakub Kicinski <kuba@kernel.org> CC: Steffen Klassert <steffen.klassert@secunet.com> CC: Herbert Xu <herbert@gondor.apana.org.au> CC: netdev@vger.kernel.org CC: intel-wired-lan@lists.osuosl.org Signed-off-by:
-
David S. Miller authored
Linus Walleij says: ==================== RTL8366RB tagging support This patch set adds DSA tagging support to the RTL8366RB DSA driver. There is a minor performance improvement in the tag parser compared to the previous patch set and the review tags have been collected. ==================== Signed-off-by: -
Linus Walleij authored
This activates the support to use the CPU tag to properly direct ingress traffic to the right port. Bit 15 in register RTL8368RB_CPU_CTRL_REG can be set to 1 to disable the insertion of the CPU tag which is what the code currently does. The bit 15 define calls this setting RTL8368RB_CPU_INSTAG which is confusing since the inverse meaning is implied: programmers may think that setting this bit to 1 will *enable* inserting the tag rather than disabling it, so rename this setting in bit 15 to RTL8368RB_CPU_NO_TAG which is more to the point. After this e.g. ping works out-of-the-box with the RTL8366RB. Cc: DENG Qingfang <dqfext@gmail.com> Cc: Mauri Sandberg <sandberg@mailfence.com> Reviewed-by:
Andrew Lunn <andrew@lunn.ch> Reviewed-by:
Linus Walleij <linus.walleij@linaro.org> Signed-off-by:
-
Linus Walleij authored
This implements the known parts of the Realtek 4 byte tag protocol version 0xA, as found in the RTL8366RB DSA switch. It is designated as protocol version 0xA as a different Realtek 4 byte tag format with protocol version 0x9 is known to exist in the Realtek RTL8306 chips. The tag and switch chip lacks public documentation, so the tag format has been reverse-engineered from packet dumps. As only ingress traffic has been available for analysis an egress tag has not been possible to develop (even using educated guesses about bit fields) so this is as far as it gets. It is not known if the switch even supports egress tagging. Excessive attempts to figure out the egress tag format was made. When nothing else worked, I just tried all bit combinations with 0xannp where a is protocol and p is port. I looped through all values several times trying to get a response from ping, without any positive result. Using just these ingress tags however, the switch functionality is vastly improved and the packets find their way into the destination port without any tricky VLAN configuration. On the D-Link DIR-685 the LAN ports now come up and respond to ping without any command line configuration so this is a real improvement for users. Egress packets need to be restricted to the proper target ports using VLAN, which the RTL8366RB DSA switch driver already sets up. Cc: DENG Qingfang <dqfext@gmail.com> Cc: Mauri Sandberg <sandberg@mailfence.com> Reviewed-by:
-
Meir Lichtinger authored
This patch exposes new link modes using 100Gbps per lane, including 100G, 200G and 400G modes. Signed-off-by:
Meir Lichtinger <meirl@mellanox.com> Reviewed-by:
Aya Levin <ayal@mellanox.com> Signed-off-by:
Saeed Mahameed <saeedm@mellanox.com> Signed-off-by:
-
Meir Lichtinger authored
Define 100G, 200G and 400G link modes using 100Gbps per lane LR, ER and FR are defined as a single link mode because they are using same technology and by design are fully interoperable. EEPROM content indicates if the module is LR, ER, or FR, and the user space ethtool decoder is planned to support decoding these modes in the EEPROM. Signed-off-by:
-
David S. Miller authored
Michael Chan says: ==================== bnxt_en: Driver update for net-next. This patchset implements ethtool -X to setup user-defined RSS indirection table. The new infrastructure also allows the proper logical ring index to be used to populate the RSS indirection when queried by ethtool -x. Prior to these patches, we were incorrectly populating the output of ethtool -x with internal ring IDs which would make no sense to the user. The last 2 patches add some cleanups to the VLAN acceleration logic and check the firmware capabilities before allowing VLAN acceleration offloads. v4: Move bnxt_get_rxfh_indir_size() fix to a new patch #2. Modify patch #7 to revert RSS map to default only when necessary. v3: Use ALIGN() in patch 5. Add warning messages in patch 6. v2: Some RSS indirection table changes requested by Jakub Kicinski. ==================== Reviewed-by:Jakub Kicinski <kuba@kernel.org> Signed-off-by:
-
Edwin Peer authored
Bare-metal use cases require giving firmware and the embedded application processor control over VLAN offloads. The driver should not attempt to override or utilize this feature in such scenarios since it will not work as expected. Signed-off-by:
Edwin Peer <edwin.peer@broadcom.com> Signed-off-by:
Michael Chan <michael.chan@broadcom.com> Signed-off-by:
-
Edwin Peer authored
The hardware VLAN offload feature on our NIC does not have separate knobs for handling customer and service tags on RX. Either offloading of both must be enabled or both must be disabled. Introduce definitions for the combined feature set in order to clean up the code and make this constraint more clear. Technically these features can be separately enabled on TX, however, since the default is to turn both on, the combined TX feature set is also introduced for code consistency. Signed-off-by:
-
Michael Chan authored
With the new infrastructure in place, we can now support the setting of the indirection table from ethtool. When changing channels, in a rare case that firmware cannot reserve the rings that were promised, we will still try to keep the RSS map and only revert to default when absolutely necessary. v4: Revert RSS map to default during ring change only when absolutely necessary. v3: Add warning messages when firmware cannot reserve the requested RX rings, and when the RSS table entries have to change to default. v2: When changing channels, if the RSS table size changes and RSS map is non-default, return error. Signed-off-by: -
Michael Chan authored
Now that we have the logical indirection table, we can return these proper logical indices directly to ethtool -x instead of the physical IDs. Reported-by:
Jakub Kicinski <kicinski@fb.com> Signed-off-by:
-
Michael Chan authored
Now that we have the logical table, we can fill the HW RSS table using the logical table's entries and converting them to the HW specific format. Re-initialize the logical table to standard distribution if the number of RX rings changes during ring reservation. v4: Use bnxt_get_rxfh_indir_size() to get the RSS table size. v2: Use ALIGN() to roundup the RSS table size. Signed-off-by:
-
Michael Chan authored
On some chips, this varies based on the number of RX rings. Add this helper function and refactor the existing code to use it. Signed-off-by:
-
Michael Chan authored
The driver currently does not keep track of the logical RSS indirection table. The hardware RSS table is set up with standard default ring distribution when initializing the chip. This makes it difficult to support user sepcified indirection table entries. As a first step, add the logical table in the main bnxt structure and allocate it according to chip specific table size. Add a function that sets up default RSS distribution based on the number of RX rings. v4: Use bnxt_get_rxfh_indir_size() for the current RSS table size. v2: Use kmalloc_array() since we init. all entries afterwards. Use ALIGN() to roundup the RSS table size. Use ethtool_rxfh_indir_default() to init. the default entries. Signed-off-by: -
Michael Chan authored
Fix up bnxt_get_rxfh_indir_size() to return the proper current RSS table size for P5 chips. Change it to non-static so that bnxt.c can use it to get the table size. Signed-off-by:
-
Michael Chan authored
Currently, we allocate one page for the hardware DMA RSS indirection table. While the size is currently big enough for all chips, future chip variations may support bigger sizes, so it is better to calculate and store the chip specific size and allocate accordingly. Signed-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf-nextDavid S. Miller authored
Pablo Neira Ayuso says: ==================== Netfilter/IPVS updates for net-next The following patchset contains Netfilter updates for net-next: 1) Support for rejecting packets from the prerouting chain, from Laura Garcia Liebana. 2) Remove useless assignment in pipapo, from Stefano Brivio. 3) On demand hook registration in IPVS, from Julian Anastasov. 4) Expire IPVS connection from process context to not overload timers, also from Julian. 5) Fallback to conntrack TCP tracker to handle connection reuse in IPVS, from Julian Anastasov. 6) Several patches to support for chain bindings. 7) Expose enum nft_chain_flags through UAPI. 8) Reject unsupported chain flags from the netlink control plane. ==================== Signed-off-by: -
David S. Miller authored
Florian Fainelli says: ==================== net: phy: Uninline PHY ethtool statistics Now that we have introduced ethtool_phy_ops we can uninline those operations and move them back into phy.c where they belong. Since those functions are used by DSA, we need to continue exporting those symbols. It might be possible to remove ndo_get_ethtool_phy_stats in a subsequent patch since we could have DSA register its own ethtool_phy_ops instance instead of overloading the ethtool_ops. ==================== Signed-off-by: -
Florian Fainelli authored
Now that we have moved the PHY ethtool statistics to be dynamically registered, we no longer need to inline those for ethtool. This used to be done to avoid cross symbol referencing and allow ethtool to be decoupled from PHYLIB entirely. Signed-off-by:
-
Florian Fainelli authored
Extend ethtool_phy_ops to include the 3 function pointers necessary for implementing PHY statistics. In a subsequent change we will uninline those functions. Signed-off-by:
-
