- 07 Oct, 2016 24 commits
-
-
Bhaktipriya Shridhar authored
alloc_ordered_workqueue() with WQ_MEM_RECLAIM set, replaces deprecated create_singlethread_workqueue(). This is the identity conversion. The workqueue "ib_nl" queues work item &ib_nl_timed_work. It has been identity converted. WQ_MEM_RECLAIM has been set to ensure forward progress under memory pressure. Signed-off-by:
Bhaktipriya Shridhar <bhaktipriya96@gmail.com> Signed-off-by:
Doug Ledford <dledford@redhat.com>
-
Aviv Heller authored
When LAG is active, QP tx affinity (the physical port to which a QP is affined, or the TIS in case of raw-eth) is set in a round robin fashion during state transition from RESET to INIT. Signed-off-by:
Aviv Heller <avivh@mellanox.com> Signed-off-by:
Leon Romanovsky <leon@kernel.org> Signed-off-by:
-
Aviv Heller authored
IB bond device name is now 'mlx5_bond_X', instead of 'mlx5_X'. Signed-off-by:
-
Aviv Heller authored
When LAG is active, port up/down events should be triggered by tracking the LAG master, and not one of the two slave netdevs. In the same manner, ib_query_port() should return the details of the LAG master. Signed-off-by:
-
Aviv Heller authored
This is done in two steps: 1) Issuing CREATE_VPORT_LAG in order to have Ethernet traffic from both ports arriving on PF0 root flowtable, so we will be able to catch all raw-eth traffic on PF0. 2) Creation of LAG demux flowtable in order to direct all non-raw-eth traffic back to its source port, assuring that normal Ethernet traffic "jumps" to the root flowtable of its RX port (non-LAG behavior). Signed-off-by:
-
Aviv Heller authored
Since ib_query_port() in RoCE returns the state of its netdev as the port state, it makes sense to propagate the port up/down events to ib_core when the netdev port state changes, instead of relying on traditional core events. This also keeps both the event and ib_query_port() synchronized. Signed-off-by:
-
Yishai Hadas authored
Track asynchronous events on a receive work queue by using the mlx5_core_create_rq_tracked API. In case a fatal error has occurred letting the IB layer know about by using the ib_wq event handler. Signed-off-by:
Yishai Hadas <yishaih@mellanox.com> Signed-off-by:
-
Maor Gottlieb authored
Add support to receive Traffic Class, specific IPv6 protocol or IPv6 flow label. Signed-off-by:
Maor Gottlieb <maorg@mellanox.com> Reviewed-by:
Sagi Grimberg <sagi@grimberg.me> Signed-off-by:
-
Maor Gottlieb authored
Add support to receive TOS or specific IPv4 protocol. Signed-off-by:
-
Maor Gottlieb authored
Add the following fields to IPv6 flow filter specification: 1. Traffic Class 2. Flow Label 3. Next Header 4. Hop Limit Signed-off-by:
-
Maor Gottlieb authored
Add the following fields to IPv4 flow filter specification: 1. Type of Service 2. Time to Live 3. Flags 4. Protocol Signed-off-by:
-
Maor Gottlieb authored
Flow steering specifications structures were implemented as in an extensible way that allows one to add new filters and new fields to existing filters. These specifications have never been extended, therefore the kernel flow specifications size and the user flow specifications size were must to be equal. In downstream patch, the IPv4 flow specifications type is extended to support TOS and TTL fields. To support an extension we change the flow specifications size condition test to be as following: * If the user flow specifications is bigger than the kernel specifications, we verify that all the bits which not in the kernel specifications are zeros and the flow is added only with the kernel specifications fields. * Otherwise, we add flow rule only with the user specifications fields. User space filters must be aligned with 32bits. Signed-off-by:
-
Maor Gottlieb authored
Add validation check that all set fields in flow specification are supported by vendor. Signed-off-by:
-
Maor Gottlieb authored
Add validation check that all set fields in flow specification are supported by vendor. Signed-off-by:
-
Maor Gottlieb authored
Add support to create sniffer rule. This rule receive all incoming and outgoing packets from the port. A user could create such rule by using IB_FLOW_ATTR_SNIFFER type. Signed-off-by:
-
Maor Gottlieb authored
Move the reference count increasing of flow table to be in create_flow_rule, it will increase the reference count for each rule creation and not for each flow. Signed-off-by:
-
Maor Gottlieb authored
Fix covertiy warning of passing "&flow_attr" to function "create_flow_rule" which uses it as an array. In addition pass flow attributes argument as const. Signed-off-by:
-
Maor Gottlieb authored
Saving the flow table priority object's pointer in the flow handle is necessary for downstream patches since the sniffer flow table isn't placed at the standard flow_db structure but in a different database. Signed-off-by:
-
Maor Gottlieb authored
Fix multicast flow rule leak on adding unicast rule failure. Fixes: 038d2ef8 ('IB/mlx5: Add flow steering support') Signed-off-by:
-
Alex Vesker authored
Counters weren't updated due to raw packet QPs' traffic since the counter-id was not associated with the QP. Added support for associating the q-counter-id with the raw packet QP. The attachment is done only when changing RQ raw packet QP state from RST to INIT in modify-RQ command. FW support is required for the above, without this support raw packet QP counters will not count. Signed-off-by:
Alex Vesker <valex@mellanox.com> Signed-off-by:
-
Alex Vesker authored
Added a struct for modifying raw QP, this will allow modifying multiple parameters in raw packet QP RQ and can also be used for SQ in the future. Signed-off-by:
-
Yishai Hadas authored
Expose RSS related capabilities on both IB and vendor channels. In addition to the IB capabilities the driver reports some extra capabilities on its vendor channel: - Bit mask of the supported types of hash functions. - Bit mask of the supported RX fields that can participate in the RX hashing. Those capabilities are applicable only when the link layer is Ethernet. Signed-off-by:
-
Yishai Hadas authored
Query RSS related attributes and return them to user-space via the extended query device uverbs command. It includes both direct ones (i.e. struct ib_uverbs_rss_caps) and max_wq_type_rq which may be used in both RSS and non RSS flows. Signed-off-by:
-
Yishai Hadas authored
Expose RSS related capabilities, it includes both direct ones (i.e. struct ib_rss_caps) and max_wq_type_rq which may be used in both RSS and non RSS flows. Specifically, supported_qpts: - QP types that support RSS on the device. max_rwq_indirection_tables: - Max number of receive work queue indirection tables that could be opened on the device. max_rwq_indirection_table_size: - Max size of a receive work queue indirection table. max_wq_type_rq: - Max number of work queues of receive type that could be opened on the device. Signed-off-by:
-
- 23 Sep, 2016 6 commits
-
-
Christoph Hellwig authored
We now only use it from ib_alloc_pd to create a local DMA lkey if the device doesn't provide one, or a global rkey if the ULP requests it. This patch removes ib_get_dma_mr and open codes the functionality in ib_alloc_pd so that we can simplify the code and prevent abuse of the functionality. As a side effect we can also simplify things by removing the valid access bit check, and the PD refcounting. In the future I hope to also remove the per-PD global MR entirely by shifting this work into the HW drivers, as one step towards avoiding the struct ib_mr overload for various different use cases. Signed-off-by:
Christoph Hellwig <hch@lst.de> Reviewed-by:
Jason Gunthorpe <jgunthorpe@obsidianresearch.com> Reviewed-by:
Steve Wise <swise@opengridcomputing.com> Signed-off-by:
-
Christoph Hellwig authored
Signed-off-by:
-
Christoph Hellwig authored
Signed-off-by:
Max Gurtovoy <maxg@mellanox.com> Signed-off-by:
-
Christoph Hellwig authored
Signed-off-by:
-
Christoph Hellwig authored
Instead of exposing ib_get_dma_mr to ULPs and letting them use it more or less unchecked, this moves the capability of creating a global rkey into the RDMA core, where it can be easily audited. It also prints a warning everytime this feature is used as well. Signed-off-by:
-
Christoph Hellwig authored
This has two reasons: a) to clearly mark that drivers don't have any business using it, and b) because we're going to use it for the (dangerous) global rkey soon, so that drivers don't create on themselves. Signed-off-by:
-
- 25 Aug, 2016 5 commits
-
-
Doug Ledford authored
-
Salil authored
This patch is meant to add support of ACPI to the Hisilicon RoCE driver. Changes done are primarily meant to detect the type and then either use DT specific or ACPI spcific functions. Where ever possible, this patch tries to make use of Unified Device Property Interface APIs to support both DT and ACPI through single interface. This patch depends upon HNS ethernet driver to Reset RoCE. This function within HNS ethernet driver has also been enhanced to support ACPI and is part of other accompanying patch with this patch-set. NOTE: The changes in this patch are done over below branch, https://github.com/dledford/linux/tree/hns-roceSigned-off-by:
Salil Mehta <salil.mehta@huawei.com> Signed-off-by:
-
Salil authored
In the Hip06 SoC, the RoCE Engine is part of the Hisilicon Network Subsystem and is dependent upon DSAF module. Therefore, certain functions like RESET are exposed through the common registers of HNS DSAF module which are memory-mapped by the HNS driver and currently can only be accessed through DT/syscon interface. This patch adds the support of ACPI to the existing RoCE reset function in the HNS driver(please refer NOTE 2). Hisilicon RoCE driver (please refer NOTE 1) shall call this reset function during probe time to reset the RoCE Engine. The HNS Reset function indirectly ends up in calling the _DSM() function part of the DSDT ACPI Table. Actual reset functionality for ACPI is implemented within the ACPI DSDT Table which also has been enhanced to support this change. Support of ACPI in the HNS RoCE driver shall be pushed through a different accompanying below patch: "IB/hns: Add support of ACPI to the Hisilicon RoCE Driver" NOTE 1: HNS RoCE driver has already been accepted by its maintainer Doug Ledford<dledford@redhat.com>. Please refer below link: https://www.spinics.net/lists/linux-rdma/msg38850.html NOTE 2: RoCE reset function patch has been accepted and now is part of the net-next: https://www.mail-archive.com/netdev@vger.kernel.org/msg123867.htmlSigned-off-by:
Yisen Zhuang <yisen.zhuang@huawei.com> Signed-off-by:
-
Doug Ledford authored
-
Doug Ledford authored
Merge tag 'shared-for-4.9-2' of git://git.kernel.org/pub/scm/linux/kernel/git/leon/linux-rdma into mlx5-shared Mellanox ConnectX-4/Connect-IB shared code (SW part) * net/mlx5: Add sniffer namespaces * net/mlx5: Introduce sniffer steering hardware capabilities * net/mlx5: Configure IB devices according to LAG state * net/mlx5: Vport LAG creation support * net/mlx5: Add LAG flow steering namespace * net/mlx5: LAG demux flow table support * net/mlx5: LAG and SRIOV cannot be used together * net/mlx5e: Avoid port remapping of mlx5e netdev TISes * net/mlx5: Get RoCE netdev * net/mlx5: Implement RoCE LAG feature * net/mlx5: Add HW interfaces used by LAG * net/mlx5: Separate query_port_proto_oper for IB and EN * net/mlx5: Expose mlx5e_link_mode * net/mlx5: Update struct mlx5_ifc_xrqc_bits * net/mlx5: Modify RQ bitmask from mlx5 ifc
-
- 23 Aug, 2016 1 commit
-
-
Doug Ledford authored
Merge tag 'shared-for-4.9-1' of git://git.kernel.org/pub/scm/linux/kernel/git/leon/linux-rdma into mlx5-shared Mellanox ConnectX-4/Connect-IB shared code (HW part) * net/mlx5: Introduce alloc_encap and dealloc_encap commands * net/mlx5: Update mlx5_ifc.h for vxlan encap/decap * net/mlx5: Enable setting minimum inline header mode for VFs * net/mlx5: Improve driver log messages * net/mlx5: Unify and improve command interface * {net,IB}/mlx5: Modify QP commands via mlx5 ifc * {net,IB}/mlx5: QP/XRCD commands via mlx5 ifc * {net,IB}/mlx5: MKey/PSV commands via mlx5 ifc * {net,IB}/mlx5: CQ commands via mlx5 ifc * net/mlx5: EQ commands via mlx5 ifc * net/mlx5: Pages management commands via mlx5 ifc * net/mlx5: MCG commands via mlx5 ifc * net/mlx5: PD and UAR commands via mlx5 ifc * net/mlx5: Access register and MAD IFC commands via mlx5 ifc * net/mlx5: Init/Teardown hca commands via mlx5 ifc
-
- 22 Aug, 2016 4 commits
-
-
Wei Yongjun authored
In case of error, the function devm_ioremap_resource() returns ERR_PTR() and never returns NULL. The NULL test in the return value check should be replaced with IS_ERR(). Signed-off-by:
Wei Yongjun <weiyj.lk@gmail.com> Signed-off-by:
-
oulijun authored
This patch added maintainers for RoCE driver. Signed-off-by:
Wei Hu <xavier.huwei@huawei.com> Signed-off-by:
Lijun Ou <oulijun@huawei.com> Signed-off-by:
-
oulijun authored
This patch added Kconfig and Makefile for building RoCE module. Signed-off-by:
Nenglong Zhao <zhaonenglong@hisilicon.com> Signed-off-by:
-
oulijun authored
These are the various new source code files for the Hisilicon RoCE driver for ARM architecture. Signed-off-by:
-
