aboutsummaryrefslogtreecommitdiff
path: root/drivers/net/ethernet
AgeCommit message (Collapse)AuthorFilesLines
2020-12-09igb: use xdp_do_flushSven Auhagen1-1/+1
Since it is a new XDP implementation change xdp_do_flush_map to xdp_do_flush. Fixes: 9cbc948b5a20 ("igb: add XDP support") Suggested-by: Maciej Fijalkowski <[email protected]> Reviewed-by: Maciej Fijalkowski <[email protected]> Acked-by: Maciej Fijalkowski <[email protected]> Signed-off-by: Sven Auhagen <[email protected]> Tested-by: Sandeep Penigalapati <[email protected]> Signed-off-by: Tony Nguyen <[email protected]>
2020-12-09igb: skb add metasize for xdpSven Auhagen1-0/+4
add metasize if it is set in xdp Fixes: 9cbc948b5a20 ("igb: add XDP support") Suggested-by: Maciej Fijalkowski <[email protected]> Reviewed-by: Maciej Fijalkowski <[email protected]> Acked-by: Maciej Fijalkowski <[email protected]> Signed-off-by: Sven Auhagen <[email protected]> Tested-by: Sandeep Penigalapati <[email protected]> Signed-off-by: Tony Nguyen <[email protected]>
2020-12-09igb: XDP extack message on errorSven Auhagen1-5/+12
Add an extack error message when the RX buffer size is too small for the frame size. Fixes: 9cbc948b5a20 ("igb: add XDP support") Suggested-by: Maciej Fijalkowski <[email protected]> Reviewed-by: Maciej Fijalkowski <[email protected]> Acked-by: Maciej Fijalkowski <[email protected]> Signed-off-by: Sven Auhagen <[email protected]> Tested-by: Sandeep Penigalapati <[email protected]> Signed-off-by: Tony Nguyen <[email protected]>
2020-12-09igb: take VLAN double header into accountSven Auhagen2-4/+8
Increase the packet header padding to include double VLAN tagging. This patch uses a macro for this. Fixes: 9cbc948b5a20 ("igb: add XDP support") Suggested-by: Maciej Fijalkowski <[email protected]> Reviewed-by: Maciej Fijalkowski <[email protected]> Acked-by: Maciej Fijalkowski <[email protected]> Signed-off-by: Sven Auhagen <[email protected]> Tested-by: Sandeep Penigalapati <[email protected]> Signed-off-by: Tony Nguyen <[email protected]>
2020-12-09igb: XDP xmit back fix error codeSven Auhagen1-1/+1
The igb XDP xmit back function should only return defined error codes. Fixes: 9cbc948b5a20 ("igb: add XDP support") Reported-by: Dan Carpenter <[email protected]> Acked-by: Maciej Fijalkowski <[email protected]> Signed-off-by: Sven Auhagen <[email protected]> Tested-by: Sandeep Penigalapati <[email protected]> Signed-off-by: Tony Nguyen <[email protected]>
2020-12-09net: ena: introduce ndo_xdp_xmit() function for XDP_REDIRECTShay Agroskin2-12/+72
This patch implements the ndo_xdp_xmit() net_device function which is called when a packet is redirected to this driver using an XDP_REDIRECT directive. The function receives an array of xdp frames that it needs to xmit. The TX queues that are used to xmit these frames are the XDP queues used by the XDP_TX flow. Therefore a lock is added to synchronize both flows (XDP_TX and XDP_REDIRECT). Signed-off-by: Shay Agroskin <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
2020-12-09net: ena: use xdp_return_frame() to free xdp framesShay Agroskin2-8/+1
XDP subsystem has a function to free XDP frames and their associated pages. Using this function would help the driver's XDP implementation to adjust to new changes in the XDP subsystem in the kernel (e.g. introduction of XDP MB). Also, remove 'xdp_rx_page' field from ena_tx_buffer struct since it is no longer used. Signed-off-by: Shay Agroskin <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
2020-12-09net: ena: introduce XDP redirect implementationShay Agroskin3-32/+53
This patch adds a partial support for the XDP_REDIRECT directive which instructs the driver to pass the packet to an interface specified by the program. The directive is passed to the driver by calling bpf_redirect() or bpf_redirect_map() functions from the eBPF program. To lay the ground for integration with the existing XDP TX implementation the patch removes the redundant page ref count increase in ena_xdp_xmit_frame() and then decrease in ena_clean_rx_irq(). Instead it only DMA unmaps descriptors for which XDP TX or REDIRECT directive was received. The XDP Redirect support is still missing .ndo_xdp_xmit function implementation, which allows to redirect packet to an ENA interface, which would be added in a later patch. Signed-off-by: Shay Agroskin <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
2020-12-09net: ena: use xdp_frame in XDP TX flowShay Agroskin1-24/+29
Rename the ena_xdp_xmit_buff() function to ena_xdp_xmit_frame() and pass it an xdp_frame struct instead of xdp_buff. This change lays the ground for XDP redirect implementation which uses xdp_frames when 'xmit'ing packets. Signed-off-by: Shay Agroskin <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
2020-12-09net: ena: aggregate stats increase into a functionShay Agroskin1-99/+68
Introduce ena_increase_stat() function to increase statistics by a certain number. The function includes the - lock aquire (on 32bit machines) - stat increase - lock release (on 32bit machines) line sequence that is ubiquitous across the driver. The function increases a single stat at a time and several stats which are increased together weren't put into a function to avoid calling the function several times for each stat which looks bad and might decrease performance. Signed-off-by: Shay Agroskin <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
2020-12-09net: ena: fix coding style nitsShay Agroskin3-5/+6
This commit fixes two nits, but it does not generate any change to binary because of the optimization of gcc. - use `count` instead of `channels->combined_count` - change return type from `int` to `bool` Also add spaces and change macro order in OR assignment to make the code easier to read. Signed-off-by: Sameeh Jubran <[email protected]> Signed-off-by: Kuniyuki Iwashima <[email protected]> Signed-off-by: Shay Agroskin <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
2020-12-09net: ena: store values in their appropriate variables typesShay Agroskin2-9/+8
This patch changes some of the variables types to match the values they hold. These wrong types fail some of our static checkers that search for accidental conversions in our driver. Signed-off-by: Ido Segev <[email protected]> Signed-off-by: Igor Chauskin <[email protected]> Signed-off-by: Shay Agroskin <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
2020-12-09net: ena: add device distinct log prefix to filesShay Agroskin5-184/+312
ENA logs are adjusted to display the full ENA representation to distinct each ENA device in case of multiple interfaces. Using netdev_err/warn and dev_info functions for logging provides uniform printing with clear distinction of the device and interface. This patch changes all printing in ena_com files to use netev_* logging functions except for messages of info level. Log functions of that level would be printed with dev_info because of the early stage they are called in when net_device struct isn't yet registered. To allow using netdev_* functions in all ena_com functions, a pointer to the net_device was added to ena_com_dev struct. The patch also adds some log messages to make driver debugging easier. Signed-off-by: Amit Bernstein <[email protected]> Signed-off-by: Shay Agroskin <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
2020-12-09net: ena: use constant value for net_device allocationShay Agroskin1-23/+23
The patch changes the maximum number of RX/TX queues it advertises to the kernel (via alloc_etherdev_mq()) from a value received from the device to a constant value which is the minimum between 128 and the number of CPUs in the system. By allocating the net_device struct with a constant number of queues, the driver is able to allocate it at a much earlier stage, before calling any ena_com functions. This would allow to make all log prints in ena_com to use netdev_* log functions instead or current pr_* ones. Note: netdev_* prints in ena_com functions that are called before net_device registration in ena_probe() might print messages that are a bit ugly (with strings like "(unnamed net_device) (uninitialized)"). However we decided to use netdev_* prints in these functions anyway, for the sake of getting better messages later, when ena_com functions are called after ena_probe() form other parts of the driver. See discussion about this decision in [1]. [1] http://www.mail-archive.com/[email protected]/msg353590.html Signed-off-by: Shay Agroskin <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
2020-12-09ice: Add space to unknown speedSimon Perron Caissy1-1/+1
Add space to the end of 'Unknown' string in order to avoid concatenation with 'bps' string when formatting netdev log message. Signed-off-by: Simon Perron Caissy <[email protected]> Tested-by: Aaron Brown <[email protected]> Signed-off-by: Tony Nguyen <[email protected]>
2020-12-09ice: join format strings to same line as ice_debugJacob Keller7-160/+80
When printing messages with ice_debug, align the printed string to the origin line of the message in order to ease debugging and tracking messages back to their source. Signed-off-by: Jacob Keller <[email protected]> Tested-by: Aaron Brown <[email protected]> Signed-off-by: Tony Nguyen <[email protected]>
2020-12-09ice: silence static analysis warningBruce Allan1-1/+1
sparse warns about cast to/from restricted types which is not an actual problem; silence the warning. Signed-off-by: Bruce Allan <[email protected]> Tested-by: Aaron Brown <[email protected]> Signed-off-by: Tony Nguyen <[email protected]>
2020-12-09ice: cleanup misleading commentBruce Allan1-3/+3
The maximum Admin Queue buffer size and NVM shadow RAM sector size are both 4 Kilobytes. Some comments refer to those as 4Kb which can be confused with 4 Kilobits. Update the comments to use the commonly used KB symbol instead. Signed-off-by: Bruce Allan <[email protected]> Tested-by: Aaron Brown <[email protected]> Signed-off-by: Tony Nguyen <[email protected]>
2020-12-09ice: Remove vlan_ena from vsi structureNick Nunley2-8/+4
vlan_ena was introduced to track whether VLAN filters are enabled on the device, but 1) checking for num_vlan > 1 already gives us this information, and is currently used in this way throughout the code 2) the logic for vlan_ena is broken when multiple VLANs are active Just remove vlan_ena and use num_vlan instead. Signed-off-by: Nick Nunley <[email protected]> Tested-by: Aaron Brown <[email protected]> Signed-off-by: Tony Nguyen <[email protected]>
2020-12-09ice: Remove gate to OROM initJeb Cramer1-26/+0
Remove the gate that prevents the OROM and netlist info from being populated. The NVM now has the appropriate section for software to reference the versioning info. Signed-off-by: Jeb Cramer <[email protected]> Tested-by: Aaron Brown <[email protected]> Signed-off-by: Tony Nguyen <[email protected]>
2020-12-09ice: Enable Support for FW Override (E82X)Jeb Cramer1-4/+0
The driver is able to override the firmware when it comes to supporting a more lenient link mode. This feature was limited to E810 devices. It is now extended to E82X devices. Signed-off-by: Jeb Cramer <[email protected]> Tested-by: Aaron Brown <[email protected]> Signed-off-by: Tony Nguyen <[email protected]>
2020-12-09ice: don't always return an error for Get PHY Abilities AQ commandPaul M Stillwell Jr1-1/+2
There are times when the driver shouldn't return an error when the Get PHY abilities AQ command (0x0600) returns an error. Instead the driver should log that the error occurred and continue on. This allows the driver to load even though the AQ command failed. The user can then later determine the reason for the failure and correct it. Signed-off-by: Paul M Stillwell Jr <[email protected]> Tested-by: Aaron Brown <[email protected]> Signed-off-by: Tony Nguyen <[email protected]>
2020-12-09ice: cleanup stack hogBruce Allan1-18/+26
In ice_flow_add_prof_sync(), struct ice_flow_prof_params has recently grown in size hogging stack space when allocated there. Hogging stack space should be avoided. Change allocation to be on the heap when needed. Signed-off-by: Bruce Allan <[email protected]> Tested-by: Harikumar Bokkena <[email protected]> Signed-off-by: Tony Nguyen <[email protected]>
2020-12-09xdp: Remove the xdp_attachment_flags_ok() callbackToke Høiland-Jørgensen2-9/+0
Since commit 7f0a838254bd ("bpf, xdp: Maintain info on attached XDP BPF programs in net_device"), the XDP program attachment info is now maintained in the core code. This interacts badly with the xdp_attachment_flags_ok() check that prevents unloading an XDP program with different load flags than it was loaded with. In practice, two kinds of failures are seen: - An XDP program loaded without specifying a mode (and which then ends up in driver mode) cannot be unloaded if the program mode is specified on unload. - The dev_xdp_uninstall() hook always calls the driver callback with the mode set to the type of the program but an empty flags argument, which means the flags_ok() check prevents the program from being removed, leading to bpf prog reference leaks. The original reason this check was added was to avoid ambiguity when multiple programs were loaded. With the way the checks are done in the core now, this is quite simple to enforce in the core code, so let's add a check there and get rid of the xdp_attachment_flags_ok() callback entirely. Fixes: 7f0a838254bd ("bpf, xdp: Maintain info on attached XDP BPF programs in net_device") Signed-off-by: Toke Høiland-Jørgensen <[email protected]> Signed-off-by: Daniel Borkmann <[email protected]> Acked-by: Jakub Kicinski <[email protected]> Link: https://lore.kernel.org/bpf/[email protected]
2020-12-08net: atheros: simplify the return expression of atl2_phy_setup_autoneg_adv()Zheng Yongjun1-7/+1
Simplify the return expression. Signed-off-by: Zheng Yongjun <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08drivers: net: qlcnic: simplify the return expression of ↵Zheng Yongjun1-6/+1
qlcnic_sriov_vf_shutdown() Simplify the return expression. Signed-off-by: Zheng Yongjun <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08drivers: net: ionic: simplify the return expression of ionic_set_rxfh()Zheng Yongjun1-6/+1
Simplify the return expression. Signed-off-by: Zheng Yongjun <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08net: ll_temac: Fix potential NULL dereference in temac_probe()Zhang Changzhong1-6/+3
platform_get_resource() may fail and in this case a NULL dereference will occur. Fix it to use devm_platform_ioremap_resource() instead of calling platform_get_resource() and devm_ioremap(). This is detected by Coccinelle semantic patch. @@ expression pdev, res, n, t, e, e1, e2; @@ res = \(platform_get_resource\|platform_get_resource_byname\)(pdev, t, n); + if (!res) + return -EINVAL; ... when != res == NULL e = devm_ioremap(e1, res->start, e2); Fixes: 8425c41d1ef7 ("net: ll_temac: Extend support to non-device-tree platforms") Signed-off-by: Zhang Changzhong <[email protected]> Acked-by: Esben Haabendal <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08gve: Add support for raw addressing in the tx pathCatherine Sullivan5-42/+185
During TX, skbs' data addresses are dma_map'ed and passed to the NIC. This means that the device can perform DMA directly from these addresses and the driver does not have to copy the buffer content into pre-allocated buffers/qpls (as in qpl mode). Reviewed-by: Yangchun Fu <[email protected]> Signed-off-by: Catherine Sullivan <[email protected]> Signed-off-by: David Awogbemila <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08gve: Rx Buffer RecyclingDavid Awogbemila2-74/+135
This patch lets the driver reuse buffers that have been freed by the networking stack. In the raw addressing case, this allows the driver avoid allocating new buffers. In the qpl case, the driver can avoid copies. This patch separates the page refcount tracking mechanism into a function gve_rx_can_recycle_buffer which uses get_page - this will be changed in a future patch to entirely eliminate the use of get_page in tracking page refcounts. Signed-off-by: David Awogbemila <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08gve: Add support for raw addressing to the rx pathCatherine Sullivan5-62/+182
Add support to use raw dma addresses in the rx path. Due to this new support we can alloc a new buffer instead of making a copy. RX buffers are handed to the networking stack and are re-allocated as needed, avoiding the need to use skb_copy_to_linear_data() as in "qpl" mode. Reviewed-by: Yangchun Fu <[email protected]> Signed-off-by: Catherine Sullivan <[email protected]> Signed-off-by: David Awogbemila <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08gve: Add support for raw addressing device optionCatherine Sullivan4-4/+92
Add support to describe device for parsing device options. As the first device option, add raw addressing. "Raw Addressing" mode (as opposed to the current "qpl" mode) is an operational mode which allows the driver avoid bounce buffer copies which it currently performs using pre-allocated qpls (queue_page_lists) when sending and receiving packets. For egress packets, the provided skb data addresses will be dma_map'ed and passed to the device, allowing the NIC can perform DMA directly - the driver will not have to copy the buffer content into pre-allocated buffers/qpls (as in qpl mode). For ingress packets, copies are also eliminated as buffers are handed to the networking stack and then recycled or re-allocated as necessary, avoiding the use of skb_copy_to_linear_data(). This patch only introduces the option to the driver. Subsequent patches will add the ingress and egress functionality. Reviewed-by: Yangchun Fu <[email protected]> Signed-off-by: Catherine Sullivan <[email protected]> Signed-off-by: David Awogbemila <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08mlxsw: spectrum_switchdev: Allow joining VxLAN to 802.1ad bridgeAmit Cohen1-2/+2
The previous patches added support for VxLAN device enslaved to 802.1ad bridge in Spectrum-2 ASIC and vetoed it in Spectrum-1. Do not veto VxLAN with 802.1ad bridge. Signed-off-by: Amit Cohen <[email protected]> Reviewed-by: Petr Machata <[email protected]> Signed-off-by: Ido Schimmel <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08mlxsw: Veto Q-in-VNI for Spectrum-1 ASICAmit Cohen3-5/+17
Implementation of Q-in-VNI is different between ASIC types, this set adds support only for Spectrum-2. Return an error when trying to create VxLAN device and enslave it to 802.1ad bridge in Spectrum-1. Signed-off-by: Amit Cohen <[email protected]> Reviewed-by: Petr Machata <[email protected]> Signed-off-by: Ido Schimmel <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08mlxsw: spectrum_switchdev: Use ops->vxlan_join() when adding VLAN to VxLAN ↵Amit Cohen1-6/+4
device Currently mlxsw_sp_switchdev_vxlan_vlan_add() always calls mlxsw_sp_bridge_8021q_vxlan_join() because VLANs were only ever added to a VLAN-filtering bridge, which is only 802.1q bridge. This set adds support for VxLAN with 802.1ad bridge, so VLAN-filtering bridge is not only 802.1q. Call ops->vxlan_join(), so mlxsw_sp_bridge_802{1q, 1ad}_vxlan_join() will be called according to bridge type. This is needed to ensure that VxLAN with 802.1ad bridge will be vetoed in Spectrum-1 with the next patch. Signed-off-by: Amit Cohen <[email protected]> Reviewed-by: Petr Machata <[email protected]> Signed-off-by: Ido Schimmel <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08mlxsw: spectrum_nve_vxlan: Add support for Q-in-VNI for Spectrum-2 ASICAmit Cohen1-0/+42
On Spectrum-2, the default setting is not to push VLAN to the decapsulated packet. This is controlled by SPVTR.ipvid_mode. Set SPVTR.ipvid_mode to always push VLAN. Without this setting, Spectrum-2 overtakes the VLAN tag of decapsulated packet for bridging. In addition, set SPVID register to use EtherType saved in mlxsw_sp_nve_config when VLAN is pushed for the NVE tunnel. Signed-off-by: Amit Cohen <[email protected]> Reviewed-by: Petr Machata <[email protected]> Signed-off-by: Ido Schimmel <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08mlxsw: spectrum: Publish mlxsw_sp_ethtype_to_sver_type()Amit Cohen2-1/+2
Declare mlxsw_sp_ethtype_to_sver_type() in spectrum.h to enable using it in other files. It will be used in the next patch to map between EtherType and the relevant value configured by SVER register. Signed-off-by: Amit Cohen <[email protected]> Reviewed-by: Petr Machata <[email protected]> Signed-off-by: Ido Schimmel <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08mlxsw: Save EtherType as part of mlxsw_sp_nve_configAmit Cohen3-4/+6
Add EtherType field to mlxsw_sp_nve_config struct. Set EtherType according to mlxsw_sp_nve_params.ethertype. Pass 'mlxsw_sp_nve_params' instead of 'mlxsw_sp_nve_params->dev' to the function which initializes mlxsw_sp_nve_config struct to know which EtherType to use. This field is needed to configure which EtherType will be used when VLAN is pushed at ingress of the tunnel port. Signed-off-by: Amit Cohen <[email protected]> Reviewed-by: Petr Machata <[email protected]> Signed-off-by: Ido Schimmel <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08mlxsw: Save EtherType as part of mlxsw_sp_nve_paramsAmit Cohen2-2/+6
Add EtherType field to mlxsw_sp_nve_params struct. Set it when VxLAN device is added to bridge device. This field is needed to configure which EtherType will be used when VLAN is pushed at ingress of the tunnel port. Use ETH_P_8021Q for tunnel port enslaved to 802.1d and 802.1q bridges. Signed-off-by: Amit Cohen <[email protected]> Reviewed-by: Petr Machata <[email protected]> Signed-off-by: Ido Schimmel <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08mlxsw: spectrum_switchdev: Create common function for joining VxLAN to ↵Amit Cohen1-3/+12
VLAN-aware bridge The code in mlxsw_sp_bridge_8021q_vxlan_join() can be used also for 802.1ad bridge. Move the code to function called mlxsw_sp_bridge_vlan_aware_vxlan_join() and call it from mlxsw_sp_bridge_8021q_vxlan_join() to enable code reuse. Signed-off-by: Amit Cohen <[email protected]> Reviewed-by: Petr Machata <[email protected]> Signed-off-by: Ido Schimmel <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08mlxsw: reg: Add support for tunnel port in SPVID registerAmit Cohen1-1/+9
Add spvid_tport field which indicates if the port is tunnel port. When spvid_tport is true, local_port field supposed to be tunnel port type. It will be used to configure which Ethertype will be used when VLAN is pushed at ingress for tunnel port. Signed-off-by: Amit Cohen <[email protected]> Reviewed-by: Petr Machata <[email protected]> Signed-off-by: Ido Schimmel <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08mlxsw: reg: Add Switch Port VLAN Stacking RegisterAmit Cohen1-0/+104
SPVTR register configures the VLAN mode of the port to enable VLAN stacking. It will be used to configure VxLAN to push VLAN to the decapsulated packet. Without this setting, Spectrum-2 overtakes the VLAN tag of decapsulated packet for bridging. Signed-off-by: Amit Cohen <[email protected]> Reviewed-by: Petr Machata <[email protected]> Signed-off-by: Ido Schimmel <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08mlxsw: Use one enum for all registers that contain tunnel_port fieldAmit Cohen3-25/+11
Currently SFN, TNUMT and TNPC registers use separate enums for tunnel_port. Create one enum with a neutral name and use it. Remove the enums that are not currently required. The next patches add two more registers that contain tunnel_port field, the new enum can be used for them also. Signed-off-by: Amit Cohen <[email protected]> Reviewed-by: Petr Machata <[email protected]> Signed-off-by: Ido Schimmel <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08Merge tag 'mlx5-updates-2020-12-01' of ↵David S. Miller27-350/+1493
git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux mlx5-updates-2020-12-01 mlx5e port TX timestamping support and MISC updates 1) Add support for port TX timestamping, for better PTP accuracy. Currently in mlx5 HW TX timestamping is done on CQE (TX completion) generation, which much earlier than when the packet actually goes out to the wire, in this series Eran implements the option to do timestamping on the port using a special SQ (Send Queue), such Send Queue will generate 2 CQEs (TX completions), the original one and a new one when the packet leaves the port, due to the nature of this special handling, such mechanism is an opt-in only and it is off by default to avoid any performance degradation on normal traffic flows. This patchset improves TX Hardware timestamping offset to be less than 40ns at a 100Gbps line rate, compared to 600ns before. With that, making our HW compliant with G.8273.2 class C, and allow Linux systems to be deployed in the 5G telco edge, where this standard is a must. 2) Misc updates and trivial improvements. Signed-off-by: David S. Miller <[email protected]>
2020-12-08net: stmmac: overwrite the dma_cap.addr64 according to HW designFugang Duan2-8/+9
The current IP register MAC_HW_Feature1[ADDR64] only defines 32/40/64 bit width, but some SOCs support others like i.MX8MP support 34 bits but it maps to 40 bits width in MAC_HW_Feature1[ADDR64]. So overwrite dma_cap.addr64 according to HW real design. Fixes: 94abdad6974a ("net: ethernet: dwmac: add ethernet glue logic for NXP imx8 chip") Signed-off-by: Fugang Duan <[email protected]> Signed-off-by: Joakim Zhang <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08net: stmmac: delete the eee_ctrl_timer after napi disabledFugang Duan1-3/+10
There have chance to re-enable the eee_ctrl_timer and fire the timer in napi callback after delete the timer in .stmmac_release(), which introduces to access eee registers in the timer function after clocks are disabled then causes system hang. Found this issue when do suspend/resume and reboot stress test. It is safe to delete the timer after napi disabled and disable lpi mode. Fixes: d765955d2ae0b ("stmmac: add the Energy Efficient Ethernet support") Signed-off-by: Fugang Duan <[email protected]> Signed-off-by: Joakim Zhang <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08net: stmmac: free tx skb buffer in stmmac_resume()Fugang Duan1-0/+14
When do suspend/resume test, there have WARN_ON() log dump from stmmac_xmit() funciton, the code logic: entry = tx_q->cur_tx; first_entry = entry; WARN_ON(tx_q->tx_skbuff[first_entry]); In normal case, tx_q->tx_skbuff[txq->cur_tx] should be NULL because the skb should be handled and freed in stmmac_tx_clean(). But stmmac_resume() reset queue parameters like below, skb buffers may not be freed. tx_q->cur_tx = 0; tx_q->dirty_tx = 0; So free tx skb buffer in stmmac_resume() to avoid warning and memory leak. log: [ 46.139824] ------------[ cut here ]------------ [ 46.144453] WARNING: CPU: 0 PID: 0 at drivers/net/ethernet/stmicro/stmmac/stmmac_main.c:3235 stmmac_xmit+0x7a0/0x9d0 [ 46.154969] Modules linked in: crct10dif_ce vvcam(O) flexcan can_dev [ 46.161328] CPU: 0 PID: 0 Comm: swapper/0 Tainted: G O 5.4.24-2.1.0+g2ad925d15481 #1 [ 46.170369] Hardware name: NXP i.MX8MPlus EVK board (DT) [ 46.175677] pstate: 80000005 (Nzcv daif -PAN -UAO) [ 46.180465] pc : stmmac_xmit+0x7a0/0x9d0 [ 46.184387] lr : dev_hard_start_xmit+0x94/0x158 [ 46.188913] sp : ffff800010003cc0 [ 46.192224] x29: ffff800010003cc0 x28: ffff000177e2a100 [ 46.197533] x27: ffff000176ef0840 x26: ffff000176ef0090 [ 46.202842] x25: 0000000000000000 x24: 0000000000000000 [ 46.208151] x23: 0000000000000003 x22: ffff8000119ddd30 [ 46.213460] x21: ffff00017636f000 x20: ffff000176ef0cc0 [ 46.218769] x19: 0000000000000003 x18: 0000000000000000 [ 46.224078] x17: 0000000000000000 x16: 0000000000000000 [ 46.229386] x15: 0000000000000079 x14: 0000000000000000 [ 46.234695] x13: 0000000000000003 x12: 0000000000000003 [ 46.240003] x11: 0000000000000010 x10: 0000000000000010 [ 46.245312] x9 : ffff00017002b140 x8 : 0000000000000000 [ 46.250621] x7 : ffff00017636f000 x6 : 0000000000000010 [ 46.255930] x5 : 0000000000000001 x4 : ffff000176ef0000 [ 46.261238] x3 : 0000000000000003 x2 : 00000000ffffffff [ 46.266547] x1 : ffff000177e2a000 x0 : 0000000000000000 [ 46.271856] Call trace: [ 46.274302] stmmac_xmit+0x7a0/0x9d0 [ 46.277874] dev_hard_start_xmit+0x94/0x158 [ 46.282056] sch_direct_xmit+0x11c/0x338 [ 46.285976] __qdisc_run+0x118/0x5f0 [ 46.289549] net_tx_action+0x110/0x198 [ 46.293297] __do_softirq+0x120/0x23c [ 46.296958] irq_exit+0xb8/0xd8 [ 46.300098] __handle_domain_irq+0x64/0xb8 [ 46.304191] gic_handle_irq+0x5c/0x148 [ 46.307936] el1_irq+0xb8/0x180 [ 46.311076] cpuidle_enter_state+0x84/0x360 [ 46.315256] cpuidle_enter+0x34/0x48 [ 46.318829] call_cpuidle+0x18/0x38 [ 46.322314] do_idle+0x1e0/0x280 [ 46.325539] cpu_startup_entry+0x24/0x40 [ 46.329460] rest_init+0xd4/0xe0 [ 46.332687] arch_call_rest_init+0xc/0x14 [ 46.336695] start_kernel+0x420/0x44c [ 46.340353] ---[ end trace bc1ee695123cbacd ]--- Fixes: 47dd7a540b8a0 ("net: add support for STMicroelectronics Ethernet controllers.") Signed-off-by: Fugang Duan <[email protected]> Signed-off-by: Joakim Zhang <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08net: stmmac: start phylink instance before stmmac_hw_setup()Fugang Duan1-8/+8
Start phylink instance and resume back the PHY to supply RX clock to MAC before MAC layer initialization by calling .stmmac_hw_setup(), since DMA reset depends on the RX clock, otherwise DMA reset cost maximum timeout value then finally timeout. Fixes: 74371272f97f ("net: stmmac: Convert to phylink and remove phylib logic") Signed-off-by: Fugang Duan <[email protected]> Signed-off-by: Joakim Zhang <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08net: stmmac: increase the timeout for dma resetFugang Duan1-1/+1
Current timeout value is not enough for gmac5 dma reset on imx8mp platform, increase the timeout range. Signed-off-by: Fugang Duan <[email protected]> Signed-off-by: Joakim Zhang <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2020-12-08net: hns3: refine the VLAN tag handle for port based VLANGuojia Liao5-17/+64
For DEVICE_VERSION_V2, the hardware only supports max two layer VLAN tags, including port based tag inserted by hardware, tag in tx buffer descriptor(get from skb->tci) and tag in packet. For transmit packet: If port based VLAN disabled, and vf driver gets a VLAN tag from skb, the VLAN tag must be filled to the Outer_VLAN_TAG field (tag near to DMAC) of tx buffer descriptor, otherwise it may be inserted after the tag in packet. If port based VLAN enabled, and vf driver gets a VLAN tag from skb, the VLAN tag must be filled to the VLAN_TAG field (tag far to DMAC) of tx buffer descriptor, otherwise it may be conflicted with port based VLAN, and raise a hardware error. For receive packet: The hardware will strip the VLAN tags and fill them in the rx buffer descriptor, no matter port based VLAN enable or not. Because port based VLAN tag is useless for stack, so vf driver needs to discard the port based VLAN tag get from rx buffer descriptor when port based VLAN enabled. So vf must know about the port based VLAN state. For DEVICE_VERSION_V3, the hardware provides some new configuration to improve it. For transmit packet: When enable tag shift mode, hardware will handle the VLAN tag in outer_VLAN_TAG field as VLAN_TAG, so it won't conflict with port based VLAN. And hardware also make sure the tag before the tag in packet. So vf driver doesn't need to specify the tag position according to the port based VLAN state anymore. For receive packet: When enable discard mode, hardware will strip and discard the port based VLAN tag, so vf driver doesn't need to identify it from rx buffer descriptor. So modify the port based VLAN configuration, simplify the process for vf handling the VLAN tag. Signed-off-by: Guojia Liao <[email protected]> Signed-off-by: Huazhong Tan <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
Powered by cgit, Gruvboxed by Blaster4385.