kernel/linux.git/include/rdma, branch v6.18.21

RDMA/core: add rdma_rw_max_sge() helper for SQ sizing

2026-02-26T22:59:27+00:00

[ Upstream commit afcae7d7b8a278a6c29e064f99e5bafd4ac1fb37 ] svc_rdma_accept() computes sc_sq_depth as the sum of rq_depth and the number of rdma_rw contexts (ctxts). This value is used to allocate the Send CQ and to initialize the sc_sq_avail credit pool. However, when the device uses memory registration for RDMA operations, rdma_rw_init_qp() inflates the QP's max_send_wr by a factor of three per context to account for REG and INV work requests. The Send CQ and credit pool remain sized for only one work request per context, causing Send Queue exhaustion under heavy NFS WRITE workloads. Introduce rdma_rw_max_sge() to compute the actual number of Send Queue entries required for a given number of rdma_rw contexts. Upper layer protocols call this helper before creating a Queue Pair so that their Send CQs and credit accounting match the QP's true capacity. Update svc_rdma_accept() to use rdma_rw_max_sge() when computing sc_sq_depth, ensuring the credit pool reflects the work requests that rdma_rw_init_qp() will reserve. Reviewed-by: Christoph Hellwig Fixes: 00bd1439f464 ("RDMA/rw: Support threshold for registration vs scattering to local pages") Signed-off-by: Chuck Lever Link: https://patch.msgid.link/20260128005400.25147-5-cel@kernel.org Signed-off-by: Leon Romanovsky Signed-off-by: Sasha Levin

RDMA/ucma: Support write an event into a CM

2025-08-13T10:16:11+00:00

Enable user-space to inject an event into a CM through it's event channel. Two new events are added and supported: RDMA_CM_EVENT_USER and RDMA_CM_EVENT_INTERNAL. With these 2 events a new event parameter "arg" is supported, which is passed from sender to receiver transparently. With this feature an application is able to write an event into a CM channel with a new user-space rdmacm API. For example thread T1 could write an event with the API: rdma_write_cm_event(cm_id, RDMA_CM_EVENT_USER, status, arg); and thread T2 could receive the event with rdma_get_cm_event(). Signed-off-by: Mark Zhang Reviewed-by: Vlad Dumitrescu Link: https://patch.msgid.link/fdf49d0b17a45933c5d8c1d90605c9447d9a3c73.1751279794.git.leonro@nvidia.com Signed-off-by: Leon Romanovsky

RDMA/cma: Support IB service record resolution

2025-08-13T10:16:00+00:00

Add new UCMA command and the corresponding CMA implementation. Userspace can send this command to request service resolution based on service name or ID. On a successful resolution, one or multiple service records are returned, the first one will be used as destination address by default. Two new CM events are added and returned to caller accordingly: - RDMA_CM_EVENT_ADDRINFO_RESOLVED: Resolve succeeded; - RDMA_CM_EVENT_ADDRINFO_ERROR: Resolve failed. Internally two new CM states are added: - RDMA_CM_ADDRINFO_QUERY: CM is in the process of IB service resolution; - RDMA_CM_ADDRINFO_RESOLVED: CM has finished the resolve process. With these new states, beside existing state transfer processes, 2 new processes are supported: 1. The default address is used: RDMA_CM_ADDR_BOUND -> RDMA_CM_ADDRINFO_QUERY -> RDMA_CM_ADDRINFO_RESOLVED -> RDMA_CM_ROUTE_QUERY 2. To use a different address: RDMA_CM_ADDR_BOUND -> RDMA_CM_ADDRINFO_QUERY-> RDMA_CM_ADDRINFO_RESOLVED -> RDMA_CM_ADDR_QUERY -> RDMA_CM_ADDR_RESOLVED -> RDMA_CM_ROUTE_QUERY In the 2nd case, resolve_addrinfo returns multiple records, a user could call rdma_resolve_addr() with the one that is not the first. Signed-off-by: Or Har-Toov Signed-off-by: Mark Zhang Reviewed-by: Vlad Dumitrescu Link: https://patch.msgid.link/b6e82ad75522a13b5efe4ff86da0e465aab04cc2.1751279794.git.leonro@nvidia.com Signed-off-by: Leon Romanovsky

RDMA/sa_query: Support IB service records resolution

2025-08-13T10:15:56+00:00

Add an SA query API ib_sa_service_rec_get() to support building and sending SA query MADs that ask for service records with a specific name or ID, and receiving and parsing responses from the SM. Signed-off-by: Or Har-Toov Signed-off-by: Mark Zhang Reviewed-by: Vlad Dumitrescu Link: https://patch.msgid.link/9af6c82f3a3a9d975115a33235fb4ffc7c8edb21.1751279793.git.leonro@nvidia.com Signed-off-by: Leon Romanovsky

IB: Extend UVERBS_METHOD_REG_MR to get DMAH

2025-07-23T05:42:11+00:00

Extend UVERBS_METHOD_REG_MR to get DMAH and pass it to all drivers. It will be used in mlx5 driver as part of the next patch from the series. Signed-off-by: Yishai Hadas Reviewed-by: Edward Srouji Link: https://patch.msgid.link/2ae1e628c0675db81f092cc00d3ad6fbf6139405.1752752567.git.leon@kernel.org Signed-off-by: Leon Romanovsky

RDMA/core: Introduce a DMAH object and its alloc/free APIs

2025-07-23T05:42:10+00:00

Introduce a new DMA handle (DMAH) object along with its corresponding allocation and deallocation APIs. This DMAH object encapsulates attributes intended for use in DMA transactions. While its initial purpose is to support TPH functionality, it is designed to be extensible for future features such as DMA PCI multipath, PCI UIO configurations, PCI traffic class selection, and more. Further details: ---------------- We ensure that a caller requesting a DMA handle for a specific CPU ID is permitted to be scheduled on it. This prevent a potential security issue where a non privilege user may trigger DMA operations toward a CPU that it's not allowed to run on. We manage reference counting for the DMAH object and its consumers (e.g., memory regions) as will be detailed in subsequent patches in the series. Signed-off-by: Yishai Hadas Reviewed-by: Edward Srouji Link: https://patch.msgid.link/2cad097e849597e49d6b61e6865dba878257f371.1752752567.git.leon@kernel.org Signed-off-by: Leon Romanovsky

RDMA/core: Add umem "is_contiguous" and "start_dma_addr" helpers

2025-07-13T08:00:34+00:00

In some cases drivers may need to check if a given umem is contiguous. Add a helper function in core code so that drivers don't need to deal with umem or scatter-gather lists structure. Additionally add a helper for getting umem's start DMA address and use it in other helper functions that open code it. Signed-off-by: Michael Margolin Link: https://patch.msgid.link/20250708202308.24783-3-mrgolin@amazon.com Reviewed-by: Jason Gunthorpe Signed-off-by: Leon Romanovsky

RDMA/uverbs: Add a common way to create CQ with umem

2025-07-13T08:00:34+00:00

Add ioctl command attributes and a common handling for the option to create CQs with memory buffers passed from userspace. When required attributes are supplied, create umem and provide it for driver's use. The extension enables creation of CQs on top of preallocated CPU virtual or device memory buffers, by supplying VA or dmabuf fd, in a common way. Drivers can support this flow by initializing a new create_cq_umem fp field in their ops struct, with a function that can handle the new parameter. Signed-off-by: Michael Margolin Link: https://patch.msgid.link/20250708202308.24783-2-mrgolin@amazon.com Reviewed-by: Jason Gunthorpe Signed-off-by: Leon Romanovsky

RDMA/uverbs: Add empty rdma_uattrs_has_raw_cap() declaration

2025-07-09T07:41:21+00:00

The call to rdma_uattrs_has_raw_cap() is placed in mlx5 fs.c file, which is compiled without relation to CONFIG_INFINIBAND_USER_ACCESS. Despite the check is used only in flows with CONFIG_INFINIBAND_USER_ACCESS=y|m, the compilers generate the following error for CONFIG_INFINIBAND_USER_ACCESS=n builds. >> ERROR: modpost: "rdma_uattrs_has_raw_cap" [drivers/infiniband/hw/mlx5/mlx5_ib.ko] undefined! Fixes: f458ccd2aa2c ("RDMA/uverbs: Check CAP_NET_RAW in user namespace for flow create") Reported-by: kernel test robot Closes: https://lore.kernel.org/oe-kbuild-all/202507080725.bh7xrhpg-lkp@intel.com/ Link: https://patch.msgid.link/72dee6b379bd709255a5d8e8010b576d50e47170.1751967071.git.leon@kernel.org Reviewed-by: Zhu Yanjun Reviewed-by: Parav Pandit Signed-off-by: Leon Romanovsky

RDMA/uverbs: Check CAP_NET_RAW in user namespace for flow create

2025-07-01T09:21:27+00:00

Currently, the capability check is done in the default init_user_ns user namespace. When a process runs in a non default user namespace, such check fails. Due to this when a process is running using Podman, it fails to create the flow resource. Since the RDMA device is a resource within a network namespace, use the network namespace associated with the RDMA device to determine its owning user namespace. Fixes: 436f2ad05a0b ("IB/core: Export ib_create/destroy_flow through uverbs") Signed-off-by: Parav Pandit Suggested-by: Eric W. Biederman Link: https://patch.msgid.link/6df6f2f24627874c4f6d041c19dc1f6f29f68f84.1750963874.git.leon@kernel.org Signed-off-by: Leon Romanovsky