This fixes a bug where the page_offset was not being considered when
building a CQ. The HW specification says it 'must be zero', so use
a variant of mlx5_umem_find_best_quantized_pgoff() with a 0 pgoff_bitmask
to force this result.
Fixes: e126ba97db ("mlx5: Add driver for Mellanox Connect-IB adapters")
Link: https://lore.kernel.org/r/20201115114311.136250-6-leon@kernel.org
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
The RQ WQ created when making a raw ethernet QP copies the PAS list from
a dummy QPC command created earlier in the flow. The WQC and QPC PAS lists
are not fully compatible as the page_offset is a different size.
Create the RQ WQ's PAS list directly and do not try to copy it from
another command structure.
Like the prior patch, this also means that badly aligned buffers were not
correctly rejected.
Link: https://lore.kernel.org/r/20201115114311.136250-4-leon@kernel.org
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
This fixes a subtle bug, the WQ mailbox has only 5 bits to describe the
page_offset, while mlx5_ib_get_buf_offset() is hard wired to only work
with 6 bit page_offsets.
Thus it did not properly reject badly aligned buffers.
Fixes: 79b20a6c30 ("IB/mlx5: Add receive Work Queue verbs")
Fixes: 0fb2ed66a1 ("IB/mlx5: Add create and destroy functionality for Raw Packet QP")
Link: https://lore.kernel.org/r/20201115114311.136250-3-leon@kernel.org
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
SRQ uses a quantized and scaled page_offset, which is another variation of
ib_umem_find_best_pgsz(). Add mlx5_umem_find_best_quantized_pgoff() to
perform this calculation for each mailbox. A macro shows how the
calculation is directly connected to the mailbox format.
This new routine replaces the limited mlx5_ib_cont_pages() and
mlx5_ib_get_buf_offset() pairing which would reject valid configurations
rather than adjust the page_size to make it work.
In turn this is much more aggressive about choosing large page sizes for
these objects and when THP is enabled it will now often find a single page
solution.
Link: https://lore.kernel.org/r/20201115114311.136250-2-leon@kernel.org
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Use netdev->tstats instead of a member of hfi1_ipoib_dev_priv for storing
a pointer to the per-cpu counters. This allows us to use core
functionality for statistics handling.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Acked-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
gcc points out a suspicious mixing of enum types in a function that
converts from MTHCA_OPCODE_* values to IB_WC_* values:
drivers/infiniband/hw/mthca/mthca_cq.c: In function 'mthca_poll_one':
drivers/infiniband/hw/mthca/mthca_cq.c:607:21: warning: implicit conversion from 'enum <anonymous>' to 'enum ib_wc_opcode' [-Wenum-conversion]
607 | entry->opcode = MTHCA_OPCODE_INVALID;
Nothing seems to ever check for MTHCA_OPCODE_INVALID again, no idea if
this is meaningful, but it seems harmless as it deals with an invalid
input.
Remove MTHCA_OPCODE_INVALID and set the ib_wc_opcode to 0xFF, which is
still bogus, but at least doesn't make compiler warnings.
Fixes: 2a4443a699 ("[PATCH] IB/mthca: fill in opcode field for send completions")
Link: https://lore.kernel.org/r/20201026211311.3887003-1-arnd@kernel.org
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Remove the ib_is_destroyable_retryable() concept.
The idea here was to allow the drivers to forcibly clean the HW object
even if they otherwise didn't want to (eg because of usecnt). This was an
attempt to clean up in a world where drivers were not allowed to fail HW
object destruction.
Now that we are going back to allowing HW objects to fail destroy this
doesn't make sense. Instead if a uobject's HW object can't be destroyed it
is left on the uobject list and it is up to uverbs_destroy_ufile_hw() to
clean it. Multiple passes over the uobject list allow hidden dependencies
to be resolved. If that fails the HW driver is broken, throw a WARN_ON and
leak the HW object memory.
All the other tricky failure paths (eg on creation error unwind) have
already been updated to this new model.
Link: https://lore.kernel.org/r/20201104144556.3809085-2-leon@kernel.org
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
The IBTA specification has new speed - NDR. That speed supports signaling
rate of 100Gb. mlx5 IB driver translates link modes reported by ConnectX
device to IB speed and width. Added translation of new 100Gb, 200Gb and
400Gb link modes to NDR IB type and width of x1, x2 or x4 respectively.
Link: https://lore.kernel.org/r/20201026133738.1340432-3-leon@kernel.org
Signed-off-by: Meir Lichtinger <meirl@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Mixing these together is just a mess, make a dedicated version,
mlx5_ib_update_mr_pas(), which directly loads the whole MTT for a non-ODP
MR.
The split out version can trivially use a simple loop with
rdma_for_each_block() which allows using the core code to compute the MR
pages and avoids seeking in the SGL list after each chunk as the
__mlx5_ib_populate_pas() call required.
Significantly speeds loading large MTTs.
Link: https://lore.kernel.org/r/20201026132314.1336717-5-leon@kernel.org
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
The memory allocation is quite complicated, and makes this function hard
to understand. Refactor things so that a function call sets up the WR, SG,
DMA mapping and buffer, further splitting that into buffer and DMA/wr.
This also slightly changes the buffer allocation logic to try an order 0
page allocation (with OOM warnings on) before going to the emergency page.
Link: https://lore.kernel.org/r/20201026132314.1336717-4-leon@kernel.org
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
For the user MR path, instead of calling this after getting the umem, call
it as part of creating the struct mlx5_ib_mr and distill its output to a
single page_shift stored inside the mr.
This avoids passing around the tuple of its output. Based on the umem and
page_shift, the output arguments can be computed using:
count == ib_umem_num_pages(mr->umem)
shift == mr->page_shift
ncont == ib_umem_num_dma_blocks(mr->umem, 1 << mr->page_shift)
order == order_base_2(ncont)
And since mr->page_shift == umem_odp->page_shift then ncont ==
ib_umem_num_dma_blocks() == ib_umem_odp_num_pages() for ODP umems.
Link: https://lore.kernel.org/r/20201026131936.1335664-5-leon@kernel.org
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
reg_pages should always contain mr->npage since when the mr is finally
de-reg'd it is always subtracted out.
If there were any error exits then mlx5_ib_rereg_user_mr() would leave the
reg_pages adjusted and this will cause it to be double subtracted
eventually.
The manipulation of reg_pages is inherently connected to the umem, so lift
it out of set_mr_fields() and only adjust it around creating/destroying a
umem.
reg_pages is only used for diagnostics in sysfs.
Fixes: 7d0cc6edcc ("IB/mlx5: Add MR cache for large UMR regions")
Link: https://lore.kernel.org/r/20201026131936.1335664-3-leon@kernel.org
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
The is only ever set to non-zero if the MR is from the cache, and if it is
cached then the order is in cached_ent->order.
Make it clearer that use_umr_mtt_update() only returns true for cached MRs
and remove the redundant data.
Link: https://lore.kernel.org/r/20201026131936.1335664-2-leon@kernel.org
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Make changes to use sysfs_emit in the RDMA code as cocci scripts can not
be written to handle _all_ the possible variants of various sprintf family
uses in sysfs show functions.
While there, make the code more legible and update its style to be more
like the typical kernel styles.
Miscellanea:
o Use intermediate pointers for dereferences
o Add and use string lookup functions
o return early when any intermediate call fails so normal return is
at the bottom of the function
o mlx4/mcg.c:sysfs_show_group: use scnprintf to format intermediate strings
Link: https://lore.kernel.org/r/f5c9e4c9d8dafca1b7b70bd597ee7f8f219c31c8.1602122880.git.joe@perches.com
Signed-off-by: Joe Perches <joe@perches.com>
Acked-by: Jack Wang <jinpu.wang@cloud.ionos.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
The doorbell needs to store PI information into QPC, so the RoCEE should
wait for the results of storing, that is, it needs two bus operations to
complete a doorbell. When ROCEE is in SDI mode, multiple doorbells may be
interlocked because the RoCEE can only handle bus operations serially. So a
flag to mark if HIP09 is working in SDI mode is added. When the SDI flag is
set, the ROCEE will ignore the PI information of the doorbell, continue to
fetch wqe and verify its validity by it's owner_bit.
Link: https://lore.kernel.org/r/1603195493-22741-1-git-send-email-liweihang@huawei.com
Signed-off-by: Lang Cheng <chenglang@huawei.com>
Signed-off-by: Weihang Li <liweihang@huawei.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Allowing userspace to invoke these commands is probably going to crash
these drivers as they are not tested and not expecting to use them on a
user object.
For example pvrdma touches cq->ring_state which is not initialized for
user QPs.
These commands are effected:
- IB_USER_VERBS_CMD_REQ_NOTIFY_CQ is ibv_cmd_req_notify_cq() in
rdma-core, only hfi1, ipath and rxe calls it.
- IB_USER_VERBS_CMD_POLL_CQ is ibv_cmd_poll_cq() in rdma-core, only
ipath and hfi1 calls it.
- IB_USER_VERBS_CMD_POST_SEND/RECV is ibv_cmd_post_send/recv() in
rdma-core, only ipath and hfi1 call them.
- IB_USER_VERBS_CMD_POST_SRQ_RECV is ibv_cmd_post_srq_recv() in
rdma-core, only ipath and hfi1 calls it.
- IB_USER_VERBS_CMD_PEEK_CQ isn't even implemented anywhere
- IB_USER_VERBS_CMD_CREATE/DESTROY_AH is ibv_cmd_create/destroy_ah() in
rdma-core, only bnxt_re, efa, hfi1, ipath, mlx5, orcrdma, and rxe call
it.
Link: https://lore.kernel.org/r/10-v1-caa70ba3d1ab+1436e-ucmd_mask_jgg@nvidia.com
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Each driver should check that the QP attrs create_flags is supported.
Unfortuantely when create_flags was added to the QP attrs the drivers were
not updated. uverbs_ex_cmd_mask was used to block it - even though kernel
drivers use these flags too.
Check that flags is zero in all drivers that don't use it, remove
IB_USER_VERBS_EX_CMD_CREATE_QP from uverbs_ex_cmd_mask. Fix the error code
to be EOPNOTSUPP.
Link: https://lore.kernel.org/r/8-v1-caa70ba3d1ab+1436e-ucmd_mask_jgg@nvidia.com
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Each driver should check that the CQ attrs is supported. Unfortuantely
when flags was added to the CQ attrs the drivers were not updated,
uverbs_ex_cmd_mask was used to block it. This was missed when create CQ
was converted to ioctl, so non-zero flags could have been passed into
drivers.
Check that flags is zero in all drivers that don't use it, remove
IB_USER_VERBS_EX_CMD_CREATE_CQ from uverbs_ex_cmd_mask.
Fixes: 41b2a71fc8 ("IB/uverbs: Move ioctl path of create_cq and destroy_cq to a new file")
Link: https://lore.kernel.org/r/7-v1-caa70ba3d1ab+1436e-ucmd_mask_jgg@nvidia.com
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
These functions all depend on the driver providing a specific op:
- REREG_MR is rereg_user_mr(). bnxt_re set this without providing the op
- ATTACH/DEATCH_MCAST is attach_mcast()/detach_mcast(). usnic set this
without providing the op
- OPEN_QP doesn't involve the driver but requires a XRCD. qedr provides
xrcd but forgot to set it, usnic doesn't provide XRCD but set it anyhow.
- OPEN/CLOSE_XRCD are the ops alloc_xrcd()/dealloc_xrcd()
- CREATE_SRQ/DESTROY_SRQ are the ops create_srq()/destroy_srq()
- QUERY/MODIFY_SRQ is op query_srq()/modify_srq(). hns sets this but
sometimes supplies a NULL op.
- RESIZE_CQ is op resize_cq(). bnxt_re sets this boes doesn't supply an op
- ALLOC/DEALLOC_MW is alloc_mw()/dealloc_mw(). cxgb4 provided an
(now deleted) implementation but no userspace
All drivers were checked that no drivers provide the op without also
setting uverbs_cmd_mask so this should have no functional change.
Link: https://lore.kernel.org/r/4-v1-caa70ba3d1ab+1436e-ucmd_mask_jgg@nvidia.com
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Since a while now the uverbs layer checks if the driver implements a
function before allowing the ucmd to proceed. This largely obsoletes the
cmd_mask stuff, but there is some tricky bits in drivers preventing it
from being removed.
Remove the easy elements of uverbs_ex_cmd_mask by pre-setting them in the
core code. These are triggered soley based on the related ops function
pointer.
query_device_ex is not triggered based on an op, but all drivers already
implement something compatible with the extension, so enable it globally
too.
Link: https://lore.kernel.org/r/2-v1-caa70ba3d1ab+1436e-ucmd_mask_jgg@nvidia.com
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
When a mlx5 core devlink instance is reloaded in different net namespace,
its associated IB device is deleted and recreated.
Example sequence is:
$ ip netns add foo
$ devlink dev reload pci/0000:00:08.0 netns foo
$ ip netns del foo
mlx5 IB device needs to attach and detach the netdevice to it through the
netdev notifier chain during load and unload sequence. A below call graph
of the unload flow.
cleanup_net()
down_read(&pernet_ops_rwsem); <- first sem acquired
ops_pre_exit_list()
pre_exit()
devlink_pernet_pre_exit()
devlink_reload()
mlx5_devlink_reload_down()
mlx5_unload_one()
[...]
mlx5_ib_remove()
mlx5_ib_unbind_slave_port()
mlx5_remove_netdev_notifier()
unregister_netdevice_notifier()
down_write(&pernet_ops_rwsem);<- recurrsive lock
Hence, when net namespace is deleted, mlx5 reload results in deadlock.
When deadlock occurs, devlink mutex is also held. This not only deadlocks
the mlx5 device under reload, but all the processes which attempt to
access unrelated devlink devices are deadlocked.
Hence, fix this by mlx5 ib driver to register for per net netdev notifier
instead of global one, which operats on the net namespace without holding
the pernet_ops_rwsem.
Fixes: 4383cfcc65 ("net/mlx5: Add devlink reload")
Link: https://lore.kernel.org/r/20201026134359.23150-1-parav@nvidia.com
Signed-off-by: Parav Pandit <parav@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Pull rdma updates from Jason Gunthorpe:
"A usual cycle for RDMA with a typical mix of driver and core subsystem
updates:
- Driver minor changes and bug fixes for mlx5, efa, rxe, vmw_pvrdma,
hns, usnic, qib, qedr, cxgb4, hns, bnxt_re
- Various rtrs fixes and updates
- Bug fix for mlx4 CM emulation for virtualization scenarios where
MRA wasn't working right
- Use tracepoints instead of pr_debug in the CM code
- Scrub the locking in ucma and cma to close more syzkaller bugs
- Use tasklet_setup in the subsystem
- Revert the idea that 'destroy' operations are not allowed to fail
at the driver level. This proved unworkable from a HW perspective.
- Revise how the umem API works so drivers make fewer mistakes using
it
- XRC support for qedr
- Convert uverbs objects RWQ and MW to new the allocation scheme
- Large queue entry sizes for hns
- Use hmm_range_fault() for mlx5 On Demand Paging
- uverbs APIs to inspect the GID table instead of sysfs
- Move some of the RDMA code for building large page SGLs into
lib/scatterlist"
* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma: (191 commits)
RDMA/ucma: Fix use after free in destroy id flow
RDMA/rxe: Handle skb_clone() failure in rxe_recv.c
RDMA/rxe: Move the definitions for rxe_av.network_type to uAPI
RDMA: Explicitly pass in the dma_device to ib_register_device
lib/scatterlist: Do not limit max_segment to PAGE_ALIGNED values
IB/mlx4: Convert rej_tmout radix-tree to XArray
RDMA/rxe: Fix bug rejecting all multicast packets
RDMA/rxe: Fix skb lifetime in rxe_rcv_mcast_pkt()
RDMA/rxe: Remove duplicate entries in struct rxe_mr
IB/hfi,rdmavt,qib,opa_vnic: Update MAINTAINERS
IB/rdmavt: Fix sizeof mismatch
MAINTAINERS: CISCO VIC LOW LATENCY NIC DRIVER
RDMA/bnxt_re: Fix sizeof mismatch for allocation of pbl_tbl.
RDMA/bnxt_re: Use rdma_umem_for_each_dma_block()
RDMA/umem: Move to allocate SG table from pages
lib/scatterlist: Add support in dynamic allocation of SG table from pages
tools/testing/scatterlist: Show errors in human readable form
tools/testing/scatterlist: Rejuvenate bit-rotten test
RDMA/ipoib: Set rtnl_link_ops for ipoib interfaces
RDMA/uverbs: Expose the new GID query API to user space
...
