linux/drivers/infiniband/Kconfig

112 lines
3.7 KiB
Plaintext
Raw Normal View History

# SPDX-License-Identifier: GPL-2.0-only
menuconfig INFINIBAND
tristate "InfiniBand support"
depends on HAS_IOMEM && HAS_DMA
depends on NET
depends on INET
depends on m || IPV6 != m
depends on !ALPHA
select IRQ_POLL
---help---
Core support for InfiniBand (IB). Make sure to also select
any protocols you wish to use as well as drivers for your
InfiniBand hardware.
if INFINIBAND
config INFINIBAND_USER_MAD
tristate "InfiniBand userspace MAD support"
depends on INFINIBAND
---help---
Userspace InfiniBand Management Datagram (MAD) support. This
is the kernel side of the userspace MAD support, which allows
userspace processes to send and receive MADs. You will also
need libibumad from rdma-core
<https://github.com/linux-rdma/rdma-core>.
config INFINIBAND_USER_ACCESS
tristate "InfiniBand userspace access (verbs and CM)"
depends on MMU
---help---
Userspace InfiniBand access support. This enables the
kernel side of userspace verbs and the userspace
communication manager (CM). This allows userspace processes
to set up connections and directly access InfiniBand
hardware for fast-path operations. You will also need
libibverbs, libibcm and a hardware driver library from
rdma-core <https://github.com/linux-rdma/rdma-core>.
config INFINIBAND_EXP_LEGACY_VERBS_NEW_UAPI
bool "Allow experimental legacy verbs in new ioctl uAPI (EXPERIMENTAL)"
depends on INFINIBAND_USER_ACCESS
---help---
IOCTL based uAPI support for Infiniband is enabled by default for
new verbs only. This allows userspace to invoke the IOCTL based uAPI
for current legacy verbs too.
IB/uverbs: Export ib_umem_get()/ib_umem_release() to modules Export ib_umem_get()/ib_umem_release() and put low-level drivers in control of when to call ib_umem_get() to pin and DMA map userspace, rather than always calling it in ib_uverbs_reg_mr() before calling the low-level driver's reg_user_mr method. Also move these functions to be in the ib_core module instead of ib_uverbs, so that driver modules using them do not depend on ib_uverbs. This has a number of advantages: - It is better design from the standpoint of making generic code a library that can be used or overridden by device-specific code as the details of specific devices dictate. - Drivers that do not need to pin userspace memory regions do not need to take the performance hit of calling ib_mem_get(). For example, although I have not tried to implement it in this patch, the ipath driver should be able to avoid pinning memory and just use copy_{to,from}_user() to access userspace memory regions. - Buffers that need special mapping treatment can be identified by the low-level driver. For example, it may be possible to solve some Altix-specific memory ordering issues with mthca CQs in userspace by mapping CQ buffers with extra flags. - Drivers that need to pin and DMA map userspace memory for things other than memory regions can use ib_umem_get() directly, instead of hacks using extra parameters to their reg_phys_mr method. For example, the mlx4 driver that is pending being merged needs to pin and DMA map QP and CQ buffers, but it does not need to create a memory key for these buffers. So the cleanest solution is for mlx4 to call ib_umem_get() in the create_qp and create_cq methods. Signed-off-by: Roland Dreier <rolandd@cisco.com>
2007-03-05 00:15:11 +00:00
config INFINIBAND_USER_MEM
bool
depends on INFINIBAND_USER_ACCESS != n
depends on MMU
IB/uverbs: Export ib_umem_get()/ib_umem_release() to modules Export ib_umem_get()/ib_umem_release() and put low-level drivers in control of when to call ib_umem_get() to pin and DMA map userspace, rather than always calling it in ib_uverbs_reg_mr() before calling the low-level driver's reg_user_mr method. Also move these functions to be in the ib_core module instead of ib_uverbs, so that driver modules using them do not depend on ib_uverbs. This has a number of advantages: - It is better design from the standpoint of making generic code a library that can be used or overridden by device-specific code as the details of specific devices dictate. - Drivers that do not need to pin userspace memory regions do not need to take the performance hit of calling ib_mem_get(). For example, although I have not tried to implement it in this patch, the ipath driver should be able to avoid pinning memory and just use copy_{to,from}_user() to access userspace memory regions. - Buffers that need special mapping treatment can be identified by the low-level driver. For example, it may be possible to solve some Altix-specific memory ordering issues with mthca CQs in userspace by mapping CQ buffers with extra flags. - Drivers that need to pin and DMA map userspace memory for things other than memory regions can use ib_umem_get() directly, instead of hacks using extra parameters to their reg_phys_mr method. For example, the mlx4 driver that is pending being merged needs to pin and DMA map QP and CQ buffers, but it does not need to create a memory key for these buffers. So the cleanest solution is for mlx4 to call ib_umem_get() in the create_qp and create_cq methods. Signed-off-by: Roland Dreier <rolandd@cisco.com>
2007-03-05 00:15:11 +00:00
default y
config INFINIBAND_ON_DEMAND_PAGING
bool "InfiniBand on-demand paging support"
depends on INFINIBAND_USER_MEM
IB/core: Implement support for MMU notifiers regarding on demand paging regions * Add an interval tree implementation for ODP umems. Create an interval tree for each ucontext (including a count of the number of ODP MRs in this context, semaphore, etc.), and register ODP umems in the interval tree. * Add MMU notifiers handling functions, using the interval tree to notify only the relevant umems and underlying MRs. * Register to receive MMU notifier events from the MM subsystem upon ODP MR registration (and unregister accordingly). * Add a completion object to synchronize the destruction of ODP umems. * Add mechanism to abort page faults when there's a concurrent invalidation. The way we synchronize between concurrent invalidations and page faults is by keeping a counter of currently running invalidations, and a sequence number that is incremented whenever an invalidation is caught. The page fault code checks the counter and also verifies that the sequence number hasn't progressed before it updates the umem's page tables. This is similar to what the kvm module does. In order to prevent the case where we register a umem in the middle of an ongoing notifier, we also keep a per ucontext counter of the total number of active mmu notifiers. We only enable new umems when all the running notifiers complete. Signed-off-by: Sagi Grimberg <sagig@mellanox.com> Signed-off-by: Shachar Raindel <raindel@mellanox.com> Signed-off-by: Haggai Eran <haggaie@mellanox.com> Signed-off-by: Yuval Dagan <yuvalda@mellanox.com> Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-12-11 15:04:18 +00:00
select MMU_NOTIFIER
default y
---help---
On demand paging support for the InfiniBand subsystem.
Together with driver support this allows registration of
memory regions without pinning their pages, fetching the
pages on demand instead.
config INFINIBAND_ADDR_TRANS
bool "RDMA/CM"
depends on INFINIBAND
default y
---help---
Support for RDMA communication manager (CM).
This allows for a generic connection abstraction over RDMA.
config INFINIBAND_ADDR_TRANS_CONFIGFS
bool
depends on INFINIBAND_ADDR_TRANS && CONFIGFS_FS && !(INFINIBAND=y && CONFIGFS_FS=m)
default y
---help---
ConfigFS support for RDMA communication manager (CM).
This allows the user to config the default GID type that the CM
uses for each device, when initiaing new connections.
if INFINIBAND_USER_ACCESS || !INFINIBAND_USER_ACCESS
source "drivers/infiniband/hw/mthca/Kconfig"
source "drivers/infiniband/hw/qib/Kconfig"
source "drivers/infiniband/hw/cxgb3/Kconfig"
source "drivers/infiniband/hw/cxgb4/Kconfig"
source "drivers/infiniband/hw/efa/Kconfig"
source "drivers/infiniband/hw/i40iw/Kconfig"
source "drivers/infiniband/hw/mlx4/Kconfig"
source "drivers/infiniband/hw/mlx5/Kconfig"
source "drivers/infiniband/hw/ocrdma/Kconfig"
IB: Add vmw_pvrdma driver This patch series adds a driver for a paravirtual RDMA device. The device is developed for VMware's Virtual Machines and allows existing RDMA applications to continue to use existing Verbs API when deployed in VMs on ESXi. We recently did a presentation in the OFA Workshop [1] regarding this device. Description and RDMA Support ============================ The virtual device is exposed as a dual function PCIe device. One part is a virtual network device (VMXNet3) which provides networking properties like MAC, IP addresses to the RDMA part of the device. The networking properties are used to register GIDs required by RDMA applications to communicate. These patches add support and the all required infrastructure for letting applications use such a device. We support the mandatory Verbs API as well as the base memory management extensions (Local Inv, Send with Inv and Fast Register Work Requests). We currently support both Reliable Connected and Unreliable Datagram QPs but do not support Shared Receive Queues (SRQs). Also, we support the following types of Work Requests: o Send/Receive (with or without Immediate Data) o RDMA Write (with or without Immediate Data) o RDMA Read o Local Invalidate o Send with Invalidate o Fast Register Work Requests This version only adds support for version 1 of RoCE. We will add RoCEv2 support in a future patch. We do support registration of both MAC-based and IP-based GIDs. I have also created a git tree for our user-level driver [2]. Testing ======= We have tested this internally for various types of Guest OS - Red Hat, Centos, Ubuntu 12.04/14.04/16.04, Oracle Enterprise Linux, SLES 12 using backported versions of this driver. The tests included several runs of the performance tests (included with OFED), Intel MPI PingPong benchmark on OpenMPI, krping for FRWRs. Mellanox has been kind enough to test the backported version of the driver internally on their hardware using a VMware provided ESX build. I have also applied and tested this with Doug's k.o/for-4.9 branch (commit 5603910b). Note, that this patch series should be applied all together. I split out the commits so that it may be easier to review. PVRDMA Resources ================ [1] OFA Workshop Presentation - https://openfabrics.org/images/eventpresos/2016presentations/102parardma.pdf [2] Libpvrdma User-level library - http://git.openfabrics.org/?p=~aditr/libpvrdma.git;a=summary Reviewed-by: Jorgen Hansen <jhansen@vmware.com> Reviewed-by: George Zhang <georgezhang@vmware.com> Reviewed-by: Aditya Sarwade <asarwade@vmware.com> Reviewed-by: Bryan Tan <bryantan@vmware.com> Reviewed-by: Leon Romanovsky <leonro@mellanox.com> Signed-off-by: Adit Ranadive <aditr@vmware.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2016-10-03 02:10:22 +00:00
source "drivers/infiniband/hw/vmw_pvrdma/Kconfig"
source "drivers/infiniband/hw/usnic/Kconfig"
source "drivers/infiniband/hw/hns/Kconfig"
source "drivers/infiniband/hw/bnxt_re/Kconfig"
source "drivers/infiniband/hw/hfi1/Kconfig"
source "drivers/infiniband/hw/qedr/Kconfig"
source "drivers/infiniband/sw/rdmavt/Kconfig"
source "drivers/infiniband/sw/rxe/Kconfig"
endif
source "drivers/infiniband/ulp/ipoib/Kconfig"
source "drivers/infiniband/ulp/srp/Kconfig"
ib_srpt: Initial SRP Target merge for v3.3-rc1 This patch adds the kernel module ib_srpt SCSI RDMA Protocol (SRP) target implementation conforming to the SRP r16a specification for the mainline drivers/target infrastructure. This driver was originally developed by Vu Pham and has been optimized by Bart Van Assche and merged into upstream LIO based on his srpt-lio-4.1 branch here: https://github.com/bvanassche/srpt-lio/commits/srpt-lio-4.1/ This updated patch also contains the following two changes from lio-core-2.6.git/master. One is to fix a bug with 1 >= task->task_sg[] chained mappings in ib_srpt, and the other to convert the configfs control plane to reference IB Port GUID and struct srpt_port directly following mainline v4.x target_core_fabric_configfs.c convertion for ib_srpt to work with rtslib/rtsadmin v2 code. These seperate patches can be found here: ib_srpt: Fix bug with chainged SGLs in srpt_map_sg_to_ib_sge http://www.risingtidesystems.com/git/?p=lio-core-2.6.git;a=commitdiff;h=ea485147563b6555a97dbf811825fbb586519252 ib_srpt: Convert se_wwn endpoint reference to struct srpt_port->port_wwn http://www.risingtidesystems.com/git/?p=lio-core-2.6.git;a=commitdiff;h=4e544a210acb227df1bb4ca5086e65bdf4e648ea This also includes the following recent v1 -> v2 review changes: ib_srpt: Fix potential out-of-bounds array access ib_srpt: Avoid failed multipart RDMA transfers ib_srpt: Fix srpt_alloc_fabric_acl failure case return value ib_srpt: Update comments to reference $driver/$port layout ib_srpt: Fix sport->port_guid formatting code ib_srpt: Remove legacy use_port_guid_in_session_name module parameter ib_srpt: Convert srp_max_rdma_size into per port configfs attribute ib_srpt: Convert srp_max_rsp_size into per port configfs attribute ib_srpt: Convert srpt_sq_size into per port configfs attribute and v2 -> v3 review changes: ib_srpt: Fix possible race with srp_sq_size in srpt_create_ch_ib ib_srpt: Fix possible race with srp_max_rsp_size in srpt_release_channel_work ib_srpt: Fix up MAX_SRPT_RDMA_SIZE define ib_srpt: Make srpt_map_sg_to_ib_sge() failure case return -EAGAIN ib_srpt: Convert port_guid to use subnet_prefix + interface_id formatting ib_srpt: Make srpt_check_stop_free return kref_put status ib_srpt: Make compilation with BUG=n proceed` ib_srpt: Use new target_core_fabric.h include ib_srpt: Check hex2bin() return code to silence build warning Cc: Bart Van Assche <bvanassche@acm.org> Cc: Roland Dreier <roland@purestorage.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Vu Pham <vu@mellanox.com> Cc: David Dillow <dillowda@ornl.gov> Signed-off-by: Nicholas A. Bellinger <nab@risingtidesystems.com>
2011-10-14 01:30:46 +00:00
source "drivers/infiniband/ulp/srpt/Kconfig"
source "drivers/infiniband/ulp/iser/Kconfig"
iser-target: Add iSCSI Extensions for RDMA (iSER) target driver This patch adds support for iSCSI Extensions for RDMA target mode, and includes CQ pooling per isert_device context distributed across multiple active iser target sessions. It also uses cmwq process context for RX / TX ib_post_cq() polling via isert_cq_desc->cq_[rx,tx]_work invoked by isert_cq_[rx,tx]_callback() hardIRQ context callbacks. v5 changes: - Use ISER_RECV_DATA_SEG_LEN instead of hardcoded value in ISER_RX_PAD_SIZE (Or) - Fix make W=1 warnings (Or) - Add missing depends on NET && INFINIBAND_ADDR_TRANS in Kconfig (Randy + Or) - Make isert_device_find_by_ib_dev() return proper ERR_PTR (Wei Yongjun) - Properly setup iscsi_np->np_sockaddr in isert_setup_np() (Shlomi + nab) - Add special case for early ISCSI_OP_SCSI_CMD exception handling (nab) v4 changes: - Mark isert_cq_rx_work as static (Or) - Drop unnecessary ib_dma_sync_single_for_cpu + ib_dma_sync_single_for_device calls for isert_cmd->sense_buf_dma from isert_put_response (Or) - Use 12288 for ISER_RX_PAD_SIZE base to save extra page per struct iser_rx_desc (Or + nab) - Drop now unnecessary isert_rx_desc usage, and convert RX users to iser_rx_desc (Or + nab) - Move isert_[alloc,free]_rx_descriptors() ahead of isert_create_device_ib_res() usage (nab) - Mark isert_cq_[rx,tx]_callback() + prototypes as static - Fix 'warning: 'ret' may be used uninitialized' warning for isert_create_device_ib_res on powerpc allmodconfig (fengguang + nab) - Fix 'warning: 'ret' may be used uninitialized' warning for isert_connect_request on i386 allyesconfig (fengguang + nab) - Fix pr_debug conversion specification in isert_rx_completion() (fengguang + nab) - Drop unnecessary isert_conn->conn_cm_id != NULL check in isert_connect_release causing the build warning: "variable dereferenced before check 'isert_conn->conn_cm_id'" - Fix isert_lid + isert_np leak in isert_setup_np failure path - Add isert_conn->conn_wait_comp_err usage in isert_free_conn() for isert_cq_comp_err completion path - Add isert_conn->logout_posted bit to determine decrement of isert_conn->post_send_buf_count from logout response completion - Always set ISER_CONN_DOWN from isert_disconnect_work() callback v3 changes: - Convert to use per isert_cq_desc->cq_[rx,tx]_work + drop tasklets (Or + nab) - Move IB_EVENT_QP_LAST_WQE_REACHED warn into correct isert_qp_event_callback (Or) - Drop unnecessary IB_ACCESS_REMOTE_* access flag usage in isert_create_device_ib_res (Or) - Add common isert_init_send_wr(), and convert isert_put_* calls (Or) - Move to verbs+core logic to single ib_isert.[c,h] (Or + nab) - Add kmem_cache isert_cmd_cache usage for descriptor allocation (nab) - Move common ib_post_send() logic used by isert_put_*() to isert_post_response() (nab) - Add isert_put_reject call in isert_response_queue() for posting ISCSI_REJECT response. (nab) - Add ISTATE_SEND_REJECT checking in isert_do_control_comp. (nab) v2 changes: - Drop unused ISERT_ADDR_ROUTE_TIMEOUT define - Add rdma_notify() call for IB_EVENT_COMM_EST in isert_qp_event_callback() - Make isert_query_device() less verbose - Drop unused RDMA_CM_EVENT_ADDR_ERROR and RDMA_CM_EVENT_ROUTE_ERROR cases from isert_cma_handler() - Drop unused rdma/ib_fmr_pool.h include - Update isert_conn_setup_qp() to assign cq based upon least used - Add isert_create_device_ib_res() to setup PD, CQs and MRs for each underlying struct ib_device, instead of using per isert_conn resources. - Add isert_free_device_ib_res() to release PD, CQs and MRs for each underlying struct ib_device. - Add isert_device_find_by_ib_dev() - Change isert_connect_request() to drop PD, CQs and MRs allocation, and use isert_device_find_by_ib_dev() instead. - Add isert_device_try_release() - Change isert_connect_release() to decrement cq_active_qps, and drop PD, CQs and MRs resource release. - Update isert_connect_release() to call isert_device_try_release() - Make isert_create_device_ib_res() determine device->cqs_used based upon num_online_cpus() - Drop misleading isert_dump_ib_wc() usage - Drop unused rdma/ib_fmr_pool.h include - Use proper xfer_len for login PDUs in isert_rx_completion() - Add isert_release_cmd() usage - Change isert_alloc_cmd() to setup iscsi_cmd.release_cmd() pointer - Change isert_put_cmd() to perform per iscsi_opcode specific release logic - Add isert_unmap_cmd() call for ISCSI_OP_SCSI_CMD from isert_put_cmd() - Change isert_send_completion() to call atomic_dec(&isert_conn->post_send_buf_count) based upon per iscsi_opcode logic - Drop ISTATE_REMOVE processing from isert_immediate_queue() - Drop ISTATE_SEND_DATAIN processing from isert_response_queue() - Drop ISTATE_SEND_STATUS processing from isert_response_queue() - Drop iscsit_transport->iscsit_unmap_cmd() and ->iscsit_free_cmd() - Convert iser_cq_tx_tasklet() to use struct isert_cq_desc pooling logic - Convert isert_cq_tx_callback() to use struct isert_cq_desc pooling logic - Convert iser_cq_rx_tasklet() to use struct isert_cq_desc pooling logic - Convert isert_cq_rx_callback() to use struct isert_cq_desc pooling logic - Add explict iscsit_stop_dataout_timer() call to isert_do_rdma_read_comp() - Use isert_get_dataout() for iscsit_transport->iscsit_get_dataout() caller - Drop ISTATE_SEND_R2T processing from isert_immediate_queue() - Drop unused rdma/ib_fmr_pool.h include - Drop isert_cmd->cmd_kref in favor of se_cmd->cmd_kref usage - Add struct isert_device in order to support multiple EQs + CQ pooling - Add struct isert_cq_desc - Drop tasklets and cqs from isert_conn - Bump ISERT_MAX_CQ to 64 - Various minor checkpatch fixes Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com> Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
2013-03-07 08:56:19 +00:00
source "drivers/infiniband/ulp/isert/Kconfig"
source "drivers/infiniband/ulp/opa_vnic/Kconfig"
endif # INFINIBAND