Skip to content

rdma-core-64.0

Latest

Choose a tag to compare

@jgunthorpe jgunthorpe released this 08 Jul 13:06
v64.0
f272237

Changes:

  • f272237 Merge pull request #1767 from qvicksilver/fix-cache-timeout-ports
  • a07758d Merge pull request #1768 from amzn/query_port_speed
  • 9143bdd Merge pull request #1748 from jpirko/wip_umem_attrs_and_cc
  • 475b17a tests: Add tests for ibv_buf provider-aware buffers
  • 3a4a255 pyverbs: Remove MREx.close() override to fix deallocation crash
  • b882686 pyverbs: Add ibv_buf provider-aware buffer support
  • 5f7a972 pyverbs: Add ParentDomain CC unprotected alloc
  • 6a0ddf1 pyverbs: Add parent domain CC unprotected flag
  • f0d3f29 pyverbs: Add CoCo DMA bounce device cap flag
  • 98d1e04 rc_pingpong: Add unprotected memory allocation for CoCo guests
See More
  • d623370 rc_pingpong: Move buffer allocation after PD setup
  • 925a34a mlx5: Pass per-buffer dmabuf UMEM attrs to kernel
  • f6bbcda mlx5: Implement ibv_alloc_buf/ibv_free_buf ops
  • 0d19256 mlx5: Add DMA-buf heap and CoCo shared allocation support
  • dd6a30a mlx5: Adopt struct ibv_buf for internal buffer allocations
  • 684944b libibverbs: Add internal DMA-buf heap allocator library
  • 97de057 libibverbs: Extend ibv_cmd_create_qp_ex2() with driver chain
  • c9e2d0b libibverbs: Add per-buffer UMEM helper
  • e808e84 libibverbs: Add CoCo DMA bounce flag
  • 3cb2879 libibverbs: Add parent domain ALLOW_CC_UNPROTECTED_ALLOC flag
  • e5c8afe libibverbs: Fall back to the write ABI in ibv_cmd_reg_mr_ex()
  • 37ca6c1 libibverbs: Add ibv_reg_buf_mr() for provider-aware buffers
  • 91f6baf libibverbs: Accept ibv_buf in ibv_reg_mr_ex()
  • 1e4eae7 libibverbs: Fall back to legacy ops in default reg_mr_ex
  • 7bfc689 libibverbs: Add ibv_alloc_buf/ibv_free_buf verbs API
  • fb6b027 libibverbs: Add dmabuf fd field to struct ibv_buf
  • 9ff1bbd libibverbs: Introduce struct ibv_buf for common buffer description
  • 2e5f54e Update kernel headers
  • 45c66ed libibverbs: Set vmr->access in ibv_cmd_reg_mr_ex()
  • df25277 efa: Add support for the query port speed verb
  • de83d1c libibnetdisc: keep cache loadable after port timeouts
  • 1d06522 Merge pull request #1759 from qvicksilver/pkey-idx
  • 1efbcbf libibmad: echo the request P_Key index on MAD responses
  • bdf0a8f infiniband-diags: add common -Y/--pkey-idx option
  • a67c2f9 Merge pull request #1764 from amzn/memset_optimization
  • e604aff Merge pull request #1763 from abhijitG-xlnx/manpage
  • 68730b5 Merge pull request #1762 from abhijitG-xlnx/td
  • 0afcf53 Merge pull request #1760 from amzn/efa4
  • 66fcb26 libibnetdisc: guard against NULL remoteport in chassis grouping
  • 73bf0cd efa: Optimize WR memset on WR creation
  • 9c61d55 ionic: Add man pages for ionic direct verbs API
  • 0a03e4c ionic: Add thread domain support for lockless operation
  • 8b1a1e9 efa: Add EFA 0xefa4 PCI ID
  • 213649e Merge pull request #1757 from hginjgerx/fix
  • 7ed105b Merge pull request #1756 from ShacharKagan/import-devx-qp-rate
  • 145d70a libhns: Inline set extend atomic seg() to improve performance
  • 595fec3 libhns: Optimize memory allocation to reduce page faults
  • c30047f tests: Add RAW QP rate limit modification test
  • bd631f2 pyverbs: Add QP rate limit modification support
  • 9e0fec2 tests: Add UMEM and DevX export/import tests
  • c7fafed pyverbs: Add DevX and UMEM export/import support
  • d0046cd tests: Add MLX5 VAR TLP alloc and import tests
  • 48357ab pyverbs: Add VAR export/import and TLP support
  • e56695b Merge pull request #1746 from hector-cao/fix-preload-build-error
  • d20d0da librdmacm/preload.c: fix duplicate LFS symbols definitions
  • aa2c433 Merge pull request #1749 from ez-lbz/fix-node-name-map-duplicate
  • a7453f8 Merge pull request #1750 from mazorasaf/increase_max_ports
  • f710865 Merge pull request #1744 from jeholza/rxe-fix-wr-set-sge-list
  • c158c93 Merge pull request #1745 from nmorey/dev/master/systemd
  • fb62230 suse: use ldconfig_scriptlets
  • bc3e02c suse: drop dependency from libibverbs to rdma-core
  • 7b1589e cbuild: bump leap to 16.0
  • 406cd2a rxe: Fix dma.length computation in wr_set_sge_list
  • 7bc02c3 libibumad: Increase max ports supported to 192
  • a4da21a Merge pull request #1752 from liuyi-thu/rxe-fix-inline-sge-length-overflow
  • b675739 Merge pull request #1751 from liuyi-thu/rxe-fix-inline-data-list-overflow
  • a189e25 Merge pull request #1747 from Flowdalic/fix-efa-verbs
  • 665294c providers/rxe: Avoid 32-bit overflow when validating inline send length
  • 1e8ed0a providers/rxe: Fix missing length accumulation in wr_set_inline_data_list
  • 70587e6 util: Fix node name map duplicate issue
  • 1596afe providers/efa: Fix build failure when LTTNG is enabled
  • 6c11a26 Merge pull request #1743 from yishaih/mlx5_misc
  • 45184d2 mlx5: Fix inline-scatter source address on 128B CQE REQ completions
  • 1c73934 Merge pull request #1741 from jpirko/wip_sf_renames
  • 8284c5a Merge pull request #1742 from hginjgerx/fix
  • be4e4bf libhns: Fix UD 0 msg length err using new post send
  • e9de4bb kernel-boot: Add naming support for sub-function (SF) RDMA devices
  • e952c49 Merge pull request #1739 from amzn/processing-hints
  • a68c2a6 Merge pull request #1727 from nmorey/dev/factory/suse-spec
  • bed66fa Merge pull request #1725 from EdwardSro/pr-rdma-topo
  • a519952 Merge pull request #1738 from selvintxavier/bug_fixes
  • b53605f Merge pull request #1737 from pentschev/cmake-allow-sysusersdir-override
  • 2fa5086 pyverbs/efa: Add processing hints support
  • e89166b providers/efa: Add processing hints work request setter
  • 240a4f7 bnxt_re/lib: Fix the num_attrs used in bnxt_re_get_toggle_mem uapi
  • bbcdc3b bnxt_re/lib: Report vendor_err as zero for all flush completions.
  • 0cbb657 bnxt_re/lib: Fix memcpy size in bnxt_re_create_qp
  • 3d65bd6 bnxt_re/lib: Fix the flush CQE handling
  • c6dc255 bnxt_re/lib: Avoid cleaning up cq twice
  • f221577 suse: switch to systemd_requires
  • ff68db8 suse: do not BuildRequire valgrind on loongarch64
  • 8b186b5 cmake: allow overriding sysusers.d install directory
  • 82ea449 Merge pull request #1640 from tatyana-en/support-for-irdma-gen3
  • 2763c6b Merge pull request #1736 from selvintxavier/init_req
  • ef0d818 bnxt_re/lib: Initialize the SRQ request structure
  • cfba725 bnxt_re/lib: Initialize the request structures
  • 6fe804e Update library version to be 64.0
  • f8f0f10 rdma_topo: Add support for inline topologies
  • 2a24ac8 rdma_topo: Run all checks, even when some fail
  • 0434340 rdma_topo: Split NVCX_Complex into abstract and concrete classes
  • 3c7134e rdma_topo: Add offline analysis support
  • e97c7dc rdma_topo: Store unparsed device data separately
  • 97d5cf7 rdma_topo: Generalize check for supported topology
  • 0e79664 providers/irdma: Add Type2 MW support
  • 6c74950 providers/irdma: Add CQE timestamp capability for GEN 3
  • 32c9f69 providers/irdma: Add SRQ support
  • c2665c2 providers/irdma: Add Atomic Operations support
  • 971d28e providers/irdma: Set GEN 3 flag in ucontext allocation request
  • 514f06a providers/irdma: Set ibv_wc opcodes in GEN 3 RQ completions
  • d986d0b providers/irdma: Flush changes for GEN 3
  • 772db3e providers/irdma: Enable 64-byte CQE support for GEN 3

This list of changes was auto generated.