- 18 Aug, 2023 1 commit
-
-
Leon Romanovsky authored
Fix the following warning reported by kbuild: drivers/infiniband/hw/irdma/verbs.c:584: warning: Function parameter or member 'udata' not described in 'irdma_setup_umode_qp' Fixes: 3a849872 ("RDMA/irdma: Allow accurate reporting on QP max send/recv WR") Link: https://lore.kernel.org/r/2c9bcd2b773c400a1699bd7973e22bfba1e4b379.1692260011.git.leonro@nvidia.comReported-by:kernel test robot <lkp@intel.com> Closes: https://lore.kernel.org/oe-kbuild-all/202308171620.m4MNACWz-lkp@intel.com/Signed-off-by:
Leon Romanovsky <leonro@nvidia.com> Acked-by:
Shiraz Saleem <shiraz.saleem@intel.com> Signed-off-by:
Jason Gunthorpe <jgg@nvidia.com>
-
- 16 Aug, 2023 2 commits
-
-
Gustavo A. R. Silva authored
Copy union directly instead of using memcpy(). Note that in this case, a direct assignment is more readable and consistent with the subsequent assignments. This addresses the following -Wstringop-overflow warning seen in s390 with defconfig: drivers/infiniband/hw/mlx4/main.c:296:33: warning: writing 16 bytes into a region of size 0 [-Wstringop-overflow=] 296 | memcpy(&port_gid_table->gids[free].gid, | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 297 | &attr->gid, sizeof(attr->gid)); | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ This helps with the ongoing efforts to globally enable -Wstringop-overflow. Link: https://github.com/KSPP/linux/issues/308Signed-off-by:Gustavo A. R. Silva <gustavoars@kernel.org> Link: https://lore.kernel.org/r/ZNvimeRAPkJ24zRG@workReviewed-by:
Kees Cook <keescook@chromium.org> Signed-off-by:
Leon Romanovsky <leon@kernel.org>
-
Shiraz Saleem authored
The driver has code blocks for kernel push WQEs but does not map the doorbell page rendering this mode non functional [1] Remove code associated with this feature from the kernel fast path as there is currently no plan of record to support this. This also address a sparse issue reported by lkp. drivers/infiniband/hw/irdma/uk.c:285:24: sparse: sparse: incorrect type in assignment (different base types) @@ expected bool [usertype] push_wqe:1 @@ got restricted __le32 [usertype] *push_db @@ drivers/infiniband/hw/irdma/uk.c:285:24: sparse: expected bool [usertype] push_wqe:1 drivers/infiniband/hw/irdma/uk.c:285:24: sparse: got restricted __le32 [usertype] *push_db drivers/infiniband/hw/irdma/uk.c:386:24: sparse: sparse: incorrect type in assignment (different base types) @@ expected bool [usertype] push_wqe:1 @@ got restricted __le32 [usertype] *push_db @@ [1] https://lore.kernel.org/linux-rdma/20230815051809.GB22185@unreal/T/#t Fixes: 272bba19 ("RDMA: Remove unnecessary ternary operators") Fixes: 551c46ed ("RDMA/irdma: Add user/kernel shared libraries") Reported-by:
-
- 15 Aug, 2023 3 commits
-
-
Saravanan Vajravel authored
Support the new verb which indicates dmabuf support. bnxt doesn't support ODP. So use the pinned version of the dmabuf APIs to enable bnxt_re devices to work as dmabuf importer. Link: https://lore.kernel.org/r/1690790473-25850-2-git-send-email-selvin.xavier@broadcom.comSigned-off-by:
Saravanan Vajravel <saravanan.vajravel@broadcom.com> Signed-off-by:
Selvin Xavier <selvin.xavier@broadcom.com> Signed-off-by:
-
Selvin Xavier authored
Syncrhonization is required to avoid simultaneous allocation of the PD. Add a new mutex lock to handle allocation from the PD table. Signed-off-by:
Kashyap Desai <kashyap.desai@broadcom.com> Signed-off-by:
-
Kashyap Desai authored
Fix the missing dbq_lock mutex initialization Fixes: 2ad4e630 ("RDMA/bnxt_re: Implement doorbell pacing algorithm") Signed-off-by:
-
- 13 Aug, 2023 3 commits
-
-
Kalesh AP authored
When the Ethernet driver does not provide the number of lanes in the __ethtool_get_link_ksettings() response, the function ib_get_width_and_speed() does not take consideration of 50G, 100G and 200G speeds while calculating the IB width and speed. Update the width and speed for the above netdev speeds. Signed-off-by:
Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by:
-
Yue Haibing authored
Commit c2261dd7 ("RDMA/device: Add ib_device_set_netdev() as an alternative to get_netdev") declared but never implemented ib_device_netdev(), remove it. Commit 922a8e9f ("RDMA: iWARP Connection Manager.") declared but never implemented iw_cm_unbind_qp() and iw_cm_get_qp(). Signed-off-by:
Yue Haibing <yuehaibing@huawei.com> Link: https://lore.kernel.org/r/20230809142718.42316-1-yuehaibing@huawei.comSigned-off-by:
-
Guoqing Jiang authored
Replace '0' with IB_SIGNAL_REQ_WR given the sq_sig_type is either IB_SIGNAL_ALL_WR or IB_SIGNAL_REQ_WR per the below. enum ib_sig_type { IB_SIGNAL_ALL_WR, IB_SIGNAL_REQ_WR }; Signed-off-by:Guoqing Jiang <guoqing.jiang@linux.dev> Link: https://lore.kernel.org/r/20230731092106.10396-1-guoqing.jiang@linux.devSigned-off-by:
-
- 08 Aug, 2023 2 commits
-
-
Yue Haibing authored
Commit c7bcb134 ("RDMA/hns: Add SRQ support for hip08 kernel mode") declared but never implemented this. Signed-off-by:
Junxian Huang <huangjunxian6@hisilicon.com> Signed-off-by:
-
Ivan Orlov authored
Now that the driver core allows for struct class to be in read-only memory, making all 'class' structures to be declared at build time placing them into read-only memory, instead of having to be dynamically allocated at load time. Cc: Jason Gunthorpe <jgg@ziepe.ca> Cc: Leon Romanovsky <leon@kernel.org> Cc: Dennis Dalessandro <dennis.dalessandro@cornelisnetworks.com> Cc: "Md. Haris Iqbal" <haris.iqbal@ionos.com> Cc: Jack Wang <jinpu.wang@ionos.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Yishai Hadas <yishaih@nvidia.com> Cc: Ivan Orlov <ivan.orlov0322@gmail.com> Cc: Benjamin Tissoires <benjamin.tissoires@redhat.com> Suggested-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by:
Ivan Orlov <ivan.orlov0322@gmail.com> Signed-off-by:
-
- 07 Aug, 2023 12 commits
-
-
Ruan Jinjie authored
The NULL initialization of the pointers assigned by kzalloc() first is not necessary, because if the kzalloc() failed, the pointers will be assigned NULL, otherwise it works as usual. so remove it. Signed-off-by:
Ruan Jinjie <ruanjinjie@huawei.com> Link: https://lore.kernel.org/r/20230804082102.3361961-1-ruanjinjie@huawei.comSigned-off-by:
-
Xiang Yang authored
smatch reports the warning below: drivers/infiniband/core/uverbs_std_types_counters.c:110 ib_uverbs_handler_UVERBS_METHOD_COUNTERS_READ() error: 'uattr' dereferencing possible ERR_PTR() The return value of uattr maybe ERR_PTR(-ENOENT), fix this by checking the value of uattr before using it. Fixes: ebb6796b ("IB/uverbs: Add read counters support") Signed-off-by:
Xiang Yang <xiangyang3@huawei.com> Link: https://lore.kernel.org/r/20230804022525.1916766-1-xiangyang3@huawei.comSigned-off-by:
-
Chengchang Tang authored
Currently, the affinity between QP cache and CQ cache is not considered when assigning QPN, it will affect the message rate of HW. Allocate QPN from QP cache with better CQ affinity to get better performance. Fixes: 71586dd2 ("RDMA/hns: Create QP with selected QPN for bank load balance") Signed-off-by:
Chengchang Tang <tangchengchang@huawei.com> Signed-off-by:
-
Junxian Huang authored
This patch fixes inaccurate error label name in init instance. Fixes: 70f92521 ("RDMA/hns: Use the reserved loopback QPs to free MR before destroying MPT") Signed-off-by:
-
Junxian Huang authored
Currently, direct wqe is not supported for wr-list. RoCE driver excludes direct wqe for wr-list by judging whether the number of wr is 1. For a wr-list where the second wr is a length-error atomic wr, the post-send driver handles the first wr and adds 1 to the wr number counter firstly. While handling the second wr, the driver finds out a length error and terminates the wr handle process, remaining the counter at 1. This causes the driver mistakenly judges there is only 1 wr and thus enters the direct wqe process, carrying the current length-error atomic wqe. This patch fixes the error by adding a judgement whether the current wr is a bad wr. If so, use the normal doorbell process but not direct wqe despite the wr number is 1. Fixes: 01584a5e ("RDMA/hns: Add support of direct wqe") Signed-off-by:
-
Chengchang Tang authored
HW supports a variety of different speed, but the current speed is fixed. The real speed should be querried from ethernet. Fixes: 9a443537 ("IB/hns: Add driver files for hns RoCE driver") Signed-off-by:
-
Kalesh AP authored
Remove unnecessary variable initializations. Signed-off-by:
-
Kalesh AP authored
Avoid memset by initializing the variables during declaration itself. Signed-off-by:
-
Kalesh AP authored
- Remove unnecessary memset by initializing the variables during declaration itself. - Arranged variable declarartion in RCT order. Signed-off-by:
-
Selvin Xavier authored
bnxt_qplib_rcfw_alloc_sbuf allocates 24 bytes and it is better to fit on stack variables. This way we can avoid unwanted kmalloc call. Call dma_alloc_coherent directly instead of wrapper bnxt_qplib_rcfw_alloc_sbuf. Also, FW expects the side buffer needs to be aligned to BNXT_QPLIB_CMDQE_UNITS(16B). So align the size to have the extra padding bytes. Signed-off-by:
Hongguang Gao <hongguang.gao@broadcom.com> Signed-off-by:
-
Kalesh AP authored
After the cited commit, BNXT_RE_FLAG_GOT_MSIX is redundant. Remove it. Fixes: 30343221 ("bnxt_en: Remove runtime interrupt vector allocation") Signed-off-by:
-
Kalesh AP authored
Driver has not accounted QP1 for virtual functions when fetching device attributes and hence max_qp count is one less than active_qp count. Fixed driver so that it counts QP1 for virtual functions as well while fetching device attributes Fixes: ccd9d0d3 ("RDMA/bnxt_re: Enable RoCE on virtual functions") Signed-off-by:
-
- 03 Aug, 2023 1 commit
-
-
Gustavo A. R. Silva authored
One-element and zero-length arrays are deprecated. So, replace one-element array in struct irdma_qvlist_info with flexible-array member. A patch for this was sent a while ago[1]. However, it seems that, at the time, the changes were partially folded[2][3], and the actual flexible-array transformation was omitted. This patch fixes that. The only binary difference seen before/after changes is shown below: | drivers/infiniband/hw/irdma/hw.o | @@ -868,7 +868,7 @@ | drivers/infiniband/hw/irdma/hw.c:484 (discriminator 2) | size += struct_size(iw_qvlist, qv_info, rf->msix_count); | 55b: imul $0x45c,%rdi,%rdi |- 562: add $0x10,%rdi |+ 562: add $0x4,%rdi which is, of course, expected as it reflects the mistake made while folding the patch I've mentioned above. Worth mentioning is the fact that with this change we save 12 bytes of memory, as can be inferred from the diff snapshot above. Notice that: $ pahole -C rdma_qv_info idrivers/infiniband/hw/irdma/hw.o struct irdma_qv_info { u32 v_idx; /* 0 4 */ u16 ceq_idx; /* 4 2 */ u16 aeq_idx; /* 6 2 */ u8 itr_idx; /* 8 1 */ /* size: 12, cachelines: 1, members: 4 */ /* padding: 3 */ /* last cacheline: 12 bytes */ }; Link: https://lore.kernel.org/linux-hardening/20210525230038.GA175516@embeddedor/ [1] Link: https://lore.kernel.org/linux-hardening/bf46b428deef4e9e89b0ea1704b1f0e5@intel.com/ [2] Link: https://lore.kernel.org/linux-rdma/20210520143809.819-1-shiraz.saleem@intel.com/T/#u [3] Fixes: 44d9e529 ("RDMA/irdma: Implement device initialization definitions") Signed-off-by:
-
- 31 Jul, 2023 11 commits
-
-
Yue Haibing authored
commit b16f8188 ("RDMA/hns: Refactor eq code for hip06") left behind hns_roce_cleanup_eq_table(). commit 773f841a ("RDMA/hns: Avoid filling sgid index when modifying QP to RTR") leave hns_get_gid_index() unused. Remove both. Link: https://lore.kernel.org/r/20230731135916.32392-1-yuehaibing@huawei.comSigned-off-by:
-
Bob Pearson authored
If a send packet is dropped by the IP layer in rxe_requester() the call to rxe_xmit_packet() can fail with err == -EAGAIN. To recover, the state of the wqe is restored to the state before the packet was sent so it can be resent. However, the routines that save and restore the state miss a significnt part of the variable state in the wqe, the dma struct which is used to process through the sge table. And, the state is not saved before the packet is built which modifies the dma struct. Under heavy stress testing with many QPs on a fast node sending large messages to a slow node dropped packets are observed and the resent packets are corrupted because the dma struct was not restored. This patch fixes this behavior and allows the test cases to succeed. Fixes: 3050b998 ("IB/rxe: Fix race condition between requester and completer") Link: https://lore.kernel.org/r/20230721200748.4604-1-rpearsonhpe@gmail.comSigned-off-by:
Bob Pearson <rpearsonhpe@gmail.com> Signed-off-by:
-
Bob Pearson authored
This patch corrects an error in rxe_modify_srq where if the caller changes the srq size the actual new value is not returned to the caller since it may be larger than what is requested. Additionally it open codes the subroutine rcv_wqe_size() which adds very little value, and makes some whitespace changes. Fixes: 8700e3e7 ("Soft RoCE driver") Link: https://lore.kernel.org/r/20230620140142.9452-1-rpearsonhpe@gmail.comSigned-off-by:
-
Bob Pearson authored
If create_qp does not fully succeed it is possible for qp cleanup code to attempt to drain the send or recv work queues before the queues have been created causing a seg fault. This patch checks to see if the queues exist before attempting to drain them. Link: https://lore.kernel.org/r/20230620135519.9365-3-rpearsonhpe@gmail.com Reported-by: syzbot+2da1965168e7dbcba136@syzkaller.appspotmail.com Closes: https://lore.kernel.org/linux-rdma/00000000000012d89205fe7cfe00@google.com/raw Fixes: 49dc9c1f ("RDMA/rxe: Cleanup reset state handling in rxe_resp.c") Fixes: fbdeb828 ("RDMA/rxe: Cleanup error state handling in rxe_comp.c") Signed-off-by:
-
Bob Pearson authored
This patch: - Moves code to initialize a qp send work queue to a subroutine named rxe_init_sq. - Moves code to initialize a qp recv work queue to a subroutine named rxe_init_rq. - Moves initialization of qp request and response packet queues ahead of work queue initialization so that cleanup of a qp if it is not fully completed can successfully attempt to drain the packet queues without a seg fault. - Makes minor whitespace cleanups. Fixes: 8700e3e7 ("Soft RoCE driver") Link: https://lore.kernel.org/r/20230620135519.9365-2-rpearsonhpe@gmail.comSigned-off-by:
Zhu Yanjun <zyjzyj2000@gmail.com> Signed-off-by:
-
Ruan Jinjie authored
There are a little ternary operators, the true or false judgment of which is unnecessary in C language semantics. Signed-off-by:
-
Shetu Ayalew authored
The rx_dct_connect counter shows the number of received connection requests for the associated DCTs. Signed-off-by:
Shetu Ayalew <shetu@nvidia.com> Reviewed-by:
Maor Gottlieb <maorg@nvidia.com> Link: https://lore.kernel.org/r/01cd24cd7f591734741309921fdc01fc770d84a8.1690121941.git.leon@kernel.orgSigned-off-by:
-
Ruan Jinjie authored
There are many pointers assigned first, which need not to be initialized, so remove the NULL assignments. Signed-off-by:
-
Yang Li authored
Remove description of @free_hwcqp in irdma_destroy_cqp(). to silence the warning: drivers/infiniband/hw/irdma/hw.c:580: warning: Excess function parameter 'free_hwcqp' description in 'irdma_destroy_cqp' Reported-by:
Abaci Robot <abaci@linux.alibaba.com> Closes: https://bugzilla.openanolis.cn/show_bug.cgi?id=6028Signed-off-by:
Yang Li <yang.lee@linux.alibaba.com> Link: https://lore.kernel.org/r/20230731015915.34867-1-yang.lee@linux.alibaba.comReviewed-by:
Randy Dunlap <rdunlap@infradead.org> Signed-off-by:
-
Bernard Metzler authored
Immediately removing the siw module after insertion may crash in siw_stop_tx_thread(), if the according thread did not yet had a chance to initialize its wait queue and siw_stop_tx_thread() tries to wakeup that thread. Initializing the threads state before spwaning it fixes it. Reported-by:
Bernard Metzler <bmt@zurich.ibm.com> Link: https://lore.kernel.org/r/20230728114418.124328-1-bmt@zurich.ibm.comTested-by:
-
Ruan Jinjie authored
Remove unnecessary variable initializations. Signed-off-by:
-
- 30 Jul, 2023 5 commits
-
-
Sindhu Devale authored
HW GEN1 and GEN2 have different min WQ sizes but they are currently set to the same value. Use a gen specific attribute min_hw_wq_size and extend ABI to pass it to user-space. Signed-off-by:
Sindhu Devale <sindhu.devale@intel.com> Signed-off-by:
-
Sindhu Devale authored
Currently the attribute cap.max_send_wr and cap.max_recv_wr sent from user-space during create QP are the provider computed SQ/RQ depth as opposed to raw values passed from application. This inhibits computation of an accurate value for max_send_wr and max_recv_wr for this QP in the kernel which matches the value returned in user create QP. Also these capabilities needs to be reported from the driver in query QP. Add support by extending the ABI to allow the raw cap.max_send_wr and cap.max_recv_wr to be passed from user-space, while keeping compatibility for the older scheme. The internal HW depth and shift needed for the WQs needs to be computed now for both kernel and user-mode QPs. Add new helpers to assist with this: irdma_uk_calc_depth_shift_sq, irdma_uk_calc_depth_shift_rq and irdma_uk_calc_depth_shift_wq. Consolidate all the user mode QP setup into a new function irdma_setup_umode_qp which keeps it with its counterpart irdma_setup_kmode_qp. Signed-off-by:
Youvaraj Sagar <youvaraj.sagar@intel.com> Signed-off-by:
-
Haoyue Xu authored
Previously, there was no way to query the number of lanes for a network card, so the same netdev_speed would result in a fixed pair of width and speed. As network card specifications become more diverse, such fixed mode is no longer suitable, so a method is needed to obtain the correct width and speed based on the number of lanes. This patch retrieves netdev lanes and speed from net_device and translates them to IB width and speed. Signed-off-by:
Haoyue Xu <xuhaoyue1@hisilicon.com> Signed-off-by:
Luoyouming <luoyouming@huawei.com> Signed-off-by:
-
Chandramohan Akula authored
Add debug counters to track the Doorbell pacing events and report the doorbell pacing debug stats. Signed-off-by:
Chandramohan Akula <chandramohan.akula@broadcom.com> Signed-off-by:
-
Chandramohan Akula authored
Add code to expose some of the HW counters related to tx/rx data and Congestion control. Signed-off-by:
-
