linux-stable.git/drivers/infiniband, branch v6.5.2 Linux kernel stable tree RDMA/bnxt_re: Initialize dpi_tbl_lock mutex 2023-08-10T19:35:54+00:00 Kashyap Desai kashyap.desai@broadcom.com 2023-08-10T04:44:37+00:00 64b632654b97319b253c2c902fe4c11349aaa70f Fix the missing dpi_tbl_lock mutex initialization. Fixes: 0ac20faf5d83 ("RDMA/bnxt_re: Reorg the bar mapping") Link: https://lore.kernel.org/r/1691642677-21369-4-git-send-email-selvin.xavier@broadcom.com Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com> Signed-off-by: Selvin Xavier <selvin.xavier@broadcom.com> Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> 
Fix the missing dpi_tbl_lock mutex initialization.

Fixes: 0ac20faf5d83 ("RDMA/bnxt_re: Reorg the bar mapping")
Link: https://lore.kernel.org/r/1691642677-21369-4-git-send-email-selvin.xavier@broadcom.com
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Signed-off-by: Selvin Xavier <selvin.xavier@broadcom.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
RDMA/bnxt_re: Fix error handling in probe failure path 2023-08-10T19:35:54+00:00 Kalesh AP kalesh-anakkur.purayil@broadcom.com 2023-08-10T04:44:36+00:00 5ac8480ae4d01f0ca5dfd561884424046df2478a During bnxt_re_dev_init(), when bnxt_re_setup_chip_ctx() fails unregister with L2 first before bailing out probe. Fixes: ae8637e13185 ("RDMA/bnxt_re: Add chip context to identify 57500 series") Link: https://lore.kernel.org/r/1691642677-21369-3-git-send-email-selvin.xavier@broadcom.com Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Selvin Xavier <selvin.xavier@broadcom.com> Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> 
During bnxt_re_dev_init(), when bnxt_re_setup_chip_ctx() fails unregister
with L2 first before bailing out probe.

Fixes: ae8637e13185 ("RDMA/bnxt_re: Add chip context to identify 57500 series")
Link: https://lore.kernel.org/r/1691642677-21369-3-git-send-email-selvin.xavier@broadcom.com
Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Signed-off-by: Selvin Xavier <selvin.xavier@broadcom.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
RDMA/bnxt_re: Properly order ib_device_unalloc() to avoid UAF 2023-08-10T19:35:54+00:00 Selvin Xavier selvin.xavier@broadcom.com 2023-08-10T04:44:35+00:00 5363fc488da579923edf6a2fdca3d3b651dd800b ib_dealloc_device() should be called only after device cleanup. Fix the dealloc sequence. Fixes: 6d758147c7b8 ("RDMA/bnxt_re: Use auxiliary driver interface") Link: https://lore.kernel.org/r/1691642677-21369-2-git-send-email-selvin.xavier@broadcom.com Signed-off-by: Selvin Xavier <selvin.xavier@broadcom.com> Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> 
ib_dealloc_device() should be called only after device cleanup.  Fix the
dealloc sequence.

Fixes: 6d758147c7b8 ("RDMA/bnxt_re: Use auxiliary driver interface")
Link: https://lore.kernel.org/r/1691642677-21369-2-git-send-email-selvin.xavier@broadcom.com
Signed-off-by: Selvin Xavier <selvin.xavier@broadcom.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
IB/hfi1: Fix possible panic during hotplug remove 2023-08-03T18:13:57+00:00 Douglas Miller doug.miller@cornelisnetworks.com 2023-08-02T17:32:41+00:00 4fdfaef71fced490835145631a795497646f4555 During hotplug remove it is possible that the update counters work might be pending, and may run after memory has been freed. Cancel the update counters work before freeing memory. Fixes: 7724105686e7 ("IB/hfi1: add driver files") Signed-off-by: Douglas Miller <doug.miller@cornelisnetworks.com> Signed-off-by: Dennis Dalessandro <dennis.dalessandro@cornelisnetworks.com> Link: https://lore.kernel.org/r/169099756100.3927190.15284930454106475280.stgit@awfm-02.cornelisnetworks.com Signed-off-by: Leon Romanovsky <leon@kernel.org> 
During hotplug remove it is possible that the update counters work
might be pending, and may run after memory has been freed.
Cancel the update counters work before freeing memory.

Fixes: 7724105686e7 ("IB/hfi1: add driver files")
Signed-off-by: Douglas Miller <doug.miller@cornelisnetworks.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@cornelisnetworks.com>
Link: https://lore.kernel.org/r/169099756100.3927190.15284930454106475280.stgit@awfm-02.cornelisnetworks.com
Signed-off-by: Leon Romanovsky <leon@kernel.org>
RDMA/umem: Set iova in ODP flow 2023-07-31T08:38:38+00:00 Michael Guralnik michaelgur@nvidia.com 2023-07-19T09:02:41+00:00 186b169cf1e4be85aa212a893ea783a543400979 Fixing the ODP registration flow to set the iova correctly. The calculation in ib_umem_num_dma_blocks() function assumes the iova of the umem is set correctly. When iova is not set, the calculation in ib_umem_num_dma_blocks() is equivalent to length/page_size, which is true only when memory is aligned. For unaligned memory, iova must be set for the ALIGN() in the ib_umem_num_dma_blocks() to take effect and return a correct value. mlx5_ib uses ib_umem_num_dma_blocks() to decide the mkey size to use for the MR. Without this fix, when registering unaligned ODP MR, a wrong size mkey might be chosen and this might cause the UMR to fail. UMR would fail over insufficient size to update the mkey translation: infiniband mlx5_0: dump_cqe:273:(pid 0): dump error cqe 00000000: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00000010: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00000020: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00000030: 00 00 00 00 0f 00 78 06 25 00 00 58 00 da ac d2 infiniband mlx5_0: mlx5_ib_post_send_wait:806:(pid 20311): reg umr failed (6) infiniband mlx5_0: pagefault_real_mr:661:(pid 20311): Failed to update mkey page tables Fixes: f0093fb1a7cb ("RDMA/mlx5: Move mlx5_ib_cont_pages() to the creation of the mlx5_ib_mr") Fixes: a665aca89a41 ("RDMA/umem: Split ib_umem_num_pages() into ib_umem_num_dma_blocks()") Signed-off-by: Artemy Kovalyov <artemyko@nvidia.com> Signed-off-by: Michael Guralnik <michaelgur@nvidia.com> Link: https://lore.kernel.org/r/3d4be7ca2155bf239dd8c00a2d25974a92c26ab8.1689757344.git.leon@kernel.org Signed-off-by: Leon Romanovsky <leon@kernel.org> 
Fixing the ODP registration flow to set the iova correctly.
The calculation in ib_umem_num_dma_blocks() function assumes the iova of
the umem is set correctly.

When iova is not set, the calculation in ib_umem_num_dma_blocks() is
equivalent to length/page_size, which is true only when memory is aligned.
For unaligned memory, iova must be set for the ALIGN() in the
ib_umem_num_dma_blocks() to take effect and return a correct value.

mlx5_ib uses ib_umem_num_dma_blocks() to decide the mkey size to use for
the MR. Without this fix, when registering unaligned ODP MR, a wrong
size mkey might be chosen and this might cause the UMR to fail.

UMR would fail over insufficient size to update the mkey translation:
infiniband mlx5_0: dump_cqe:273:(pid 0): dump error cqe
00000000: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00000010: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00000020: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00000030: 00 00 00 00 0f 00 78 06 25 00 00 58 00 da ac d2
infiniband mlx5_0: mlx5_ib_post_send_wait:806:(pid 20311): reg umr
failed (6)
infiniband mlx5_0: pagefault_real_mr:661:(pid 20311): Failed to update
mkey page tables

Fixes: f0093fb1a7cb ("RDMA/mlx5: Move mlx5_ib_cont_pages() to the creation of the mlx5_ib_mr")
Fixes: a665aca89a41 ("RDMA/umem: Split ib_umem_num_pages() into ib_umem_num_dma_blocks()")
Signed-off-by: Artemy Kovalyov <artemyko@nvidia.com>
Signed-off-by: Michael Guralnik <michaelgur@nvidia.com>
Link: https://lore.kernel.org/r/3d4be7ca2155bf239dd8c00a2d25974a92c26ab8.1689757344.git.leon@kernel.org
Signed-off-by: Leon Romanovsky <leon@kernel.org>
RDMA/irdma: Report correct WC error 2023-07-26T11:58:42+00:00 Sindhu Devale sindhu.devale@intel.com 2023-07-25T15:54:38+00:00 ae463563b7a1b7d4a3d0b065b09d37a76b693937 Report the correct WC error if a MW bind is performed on an already valid/bound window. Fixes: 44d9e52977a1 ("RDMA/irdma: Implement device initialization definitions") Signed-off-by: Sindhu Devale <sindhu.devale@intel.com> Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com> Link: https://lore.kernel.org/r/20230725155439.1057-2-shiraz.saleem@intel.com Signed-off-by: Leon Romanovsky <leon@kernel.org> 
Report the correct WC error if a MW bind is performed
on an already valid/bound window.

Fixes: 44d9e52977a1 ("RDMA/irdma: Implement device initialization definitions")
Signed-off-by: Sindhu Devale <sindhu.devale@intel.com>
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Link: https://lore.kernel.org/r/20230725155439.1057-2-shiraz.saleem@intel.com
Signed-off-by: Leon Romanovsky <leon@kernel.org>
RDMA/irdma: Fix op_type reporting in CQEs 2023-07-26T11:58:42+00:00 Sindhu Devale sindhu.devale@intel.com 2023-07-25T15:54:37+00:00 3bfb25fa2b5bb9c29681e6ac861808f4be1331a9 The op_type field CQ poll info structure is incorrectly filled in with the queue type as opposed to the op_type received in the CQEs. The wrong opcode could be decoded and returned to the ULP. Copy the op_type field received in the CQE in the CQ poll info structure. Fixes: 24419777e943 ("RDMA/irdma: Fix RQ completion opcode") Signed-off-by: Sindhu Devale <sindhu.devale@intel.com> Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com> Link: https://lore.kernel.org/r/20230725155439.1057-1-shiraz.saleem@intel.com Signed-off-by: Leon Romanovsky <leon@kernel.org> 
The op_type field CQ poll info structure is incorrectly
filled in with the queue type as opposed to the op_type
received in the CQEs. The wrong opcode could be decoded
and returned to the ULP.

Copy the op_type field received in the CQE in the CQ poll
info structure.

Fixes: 24419777e943 ("RDMA/irdma: Fix RQ completion opcode")
Signed-off-by: Sindhu Devale <sindhu.devale@intel.com>
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Link: https://lore.kernel.org/r/20230725155439.1057-1-shiraz.saleem@intel.com
Signed-off-by: Leon Romanovsky <leon@kernel.org>
RDMA/rxe: Fix an error handling path in rxe_bind_mw() 2023-07-18T12:22:43+00:00 Christophe JAILLET christophe.jaillet@wanadoo.fr 2023-07-17T19:55:56+00:00 5c719d7aef298e9b727f39b45e88528a96df3620 All errors go to the error handling path, except this one. Be consistent and also branch to it. Fixes: 02ed253770fb ("RDMA/rxe: Introduce rxe access supported flags") Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr> Link: https://lore.kernel.org/r/43698d8a3ed4e720899eadac887427f73d7ec2eb.1689623735.git.christophe.jaillet@wanadoo.fr Reviewed-by: Bob Pearson <rpearsonhpe@gmail.com> Acked-by: Zhu Yanjun <zyjzyj2000@gmail.com> Signed-off-by: Leon Romanovsky <leon@kernel.org> 
All errors go to the error handling path, except this one. Be consistent
and also branch to it.

Fixes: 02ed253770fb ("RDMA/rxe: Introduce rxe access supported flags")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Link: https://lore.kernel.org/r/43698d8a3ed4e720899eadac887427f73d7ec2eb.1689623735.git.christophe.jaillet@wanadoo.fr
Reviewed-by: Bob Pearson <rpearsonhpe@gmail.com>
Acked-by: Zhu Yanjun <zyjzyj2000@gmail.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>
RDMA/bnxt_re: Fix hang during driver unload 2023-07-17T05:02:13+00:00 Selvin Xavier selvin.xavier@broadcom.com 2023-07-14T08:22:49+00:00 29900bf351e1a7e4643da5c3c3cd9df75c577b88 Driver unload hits a hang during stress testing of load/unload. stack trace snippet - tasklet_kill at ffffffff9aabb8b2 bnxt_qplib_nq_stop_irq at ffffffffc0a805fb [bnxt_re] bnxt_qplib_disable_nq at ffffffffc0a80c5b [bnxt_re] bnxt_re_dev_uninit at ffffffffc0a67d15 [bnxt_re] bnxt_re_remove_device at ffffffffc0a6af1d [bnxt_re] tasklet_kill can hang if the tasklet is scheduled after it is disabled. Modified the sequences to disable the interrupt first and synchronize irq before disabling the tasklet. Fixes: 1ac5a4047975 ("RDMA/bnxt_re: Add bnxt_re RoCE driver") Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com> Signed-off-by: Selvin Xavier <selvin.xavier@broadcom.com> Link: https://lore.kernel.org/r/1689322969-25402-3-git-send-email-selvin.xavier@broadcom.com Signed-off-by: Leon Romanovsky <leon@kernel.org> 
Driver unload hits a hang during stress testing of load/unload.

stack trace snippet -

tasklet_kill at ffffffff9aabb8b2
bnxt_qplib_nq_stop_irq at ffffffffc0a805fb [bnxt_re]
bnxt_qplib_disable_nq at ffffffffc0a80c5b [bnxt_re]
bnxt_re_dev_uninit at ffffffffc0a67d15 [bnxt_re]
bnxt_re_remove_device at ffffffffc0a6af1d [bnxt_re]

tasklet_kill can hang if the tasklet is scheduled after it is disabled.

Modified the sequences to disable the interrupt first and synchronize
irq before disabling the tasklet.

Fixes: 1ac5a4047975 ("RDMA/bnxt_re: Add bnxt_re RoCE driver")
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Signed-off-by: Selvin Xavier <selvin.xavier@broadcom.com>
Link: https://lore.kernel.org/r/1689322969-25402-3-git-send-email-selvin.xavier@broadcom.com
Signed-off-by: Leon Romanovsky <leon@kernel.org>
RDMA/bnxt_re: Prevent handling any completions after qp destroy 2023-07-17T05:02:13+00:00 Kashyap Desai kashyap.desai@broadcom.com 2023-07-14T08:22:48+00:00 b5bbc6551297447d3cca55cf907079e206e9cd82 HW may generate completions that indicates QP is destroyed. Driver should not be scheduling any more completion handlers for this QP, after the QP is destroyed. Since CQs are active during the QP destroy, driver may still schedule completion handlers. This can cause a race where the destroy_cq and poll_cq running simultaneously. Snippet of kernel panic while doing bnxt_re driver load unload in loop. This indicates a poll after the CQ is freed.  [77786.481636] Call Trace: [77786.481640]  <TASK> [77786.481644]  bnxt_re_poll_cq+0x14a/0x620 [bnxt_re] [77786.481658]  ? kvm_clock_read+0x14/0x30 [77786.481693]  __ib_process_cq+0x57/0x190 [ib_core] [77786.481728]  ib_cq_poll_work+0x26/0x80 [ib_core] [77786.481761]  process_one_work+0x1e5/0x3f0 [77786.481768]  worker_thread+0x50/0x3a0 [77786.481785]  ? __pfx_worker_thread+0x10/0x10 [77786.481790]  kthread+0xe2/0x110 [77786.481794]  ? __pfx_kthread+0x10/0x10 [77786.481797]  ret_from_fork+0x2c/0x50 To avoid this, complete all completion handlers before returning the destroy QP. If free_cq is called soon after destroy_qp, IB stack will cancel the CQ work before invoking the destroy_cq verb and this will prevent any race mentioned. Fixes: 1ac5a4047975 ("RDMA/bnxt_re: Add bnxt_re RoCE driver") Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com> Signed-off-by: Selvin Xavier <selvin.xavier@broadcom.com> Link: https://lore.kernel.org/r/1689322969-25402-2-git-send-email-selvin.xavier@broadcom.com Signed-off-by: Leon Romanovsky <leon@kernel.org> 
HW may generate completions that indicates QP is destroyed.
Driver should not be scheduling any more completion handlers
for this QP, after the QP is destroyed. Since CQs are active
during the QP destroy, driver may still schedule completion
handlers. This can cause a race where the destroy_cq and poll_cq
running simultaneously.

Snippet of kernel panic while doing bnxt_re driver load unload in loop.
This indicates a poll after the CQ is freed. 

[77786.481636] Call Trace:
[77786.481640]  <TASK>
[77786.481644]  bnxt_re_poll_cq+0x14a/0x620 [bnxt_re]
[77786.481658]  ? kvm_clock_read+0x14/0x30
[77786.481693]  __ib_process_cq+0x57/0x190 [ib_core]
[77786.481728]  ib_cq_poll_work+0x26/0x80 [ib_core]
[77786.481761]  process_one_work+0x1e5/0x3f0
[77786.481768]  worker_thread+0x50/0x3a0
[77786.481785]  ? __pfx_worker_thread+0x10/0x10
[77786.481790]  kthread+0xe2/0x110
[77786.481794]  ? __pfx_kthread+0x10/0x10
[77786.481797]  ret_from_fork+0x2c/0x50

To avoid this, complete all completion handlers before returning the
destroy QP. If free_cq is called soon after destroy_qp,  IB stack
will cancel the CQ work before invoking the destroy_cq verb and
this will prevent any race mentioned.

Fixes: 1ac5a4047975 ("RDMA/bnxt_re: Add bnxt_re RoCE driver")
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Signed-off-by: Selvin Xavier <selvin.xavier@broadcom.com>
Link: https://lore.kernel.org/r/1689322969-25402-2-git-send-email-selvin.xavier@broadcom.com
Signed-off-by: Leon Romanovsky <leon@kernel.org>